A. General Type- Fzzy Clsterng There are two knds of type- fzzy sets whch are often sed n clsterng algorthms: 1) nterval and ) general. In nterval typ
|
|
- Cody Dennis
- 6 years ago
- Views:
Transcription
1 014 IEEE Internatonal Conference on Fzzy Systems (FUZZ-IEEE) Jly 6-11, 014, Beng, Chna A Hybrd Type- Fzzy Clsterng Technqe for Inpt Data Preprocessng of Classfcaton Algorthms Vahd Nor, Mohammad-. Akbarzadeh-T. Departments of Compter Engneerng, Islamc Azad Unversty, Mashhad Branch, Mashhad, Iran Emals: vahd.nor@mshda.ac.r, akbarzadeh@eee.org Abstract ecently, clsterng has been sed for preprocessng datasets before applyng classfcaton algorthms n order to enhance classfcaton effcency. A strong clstered dataset as npt to classfcaton algorthms can sgnfcantly mprove comptaton tme. Ths can be partclarly sefl n Bg Data where comptaton tme s eqally or more mportant than accracy. However, there s a trade-off between speed and accracy among clsterng algorthms. Specfcally, general type- fzzy c-means (GT FCM) s consdered to be a hghly accrate clsterng approach, bt t s comptatonally ntensve. To mprove ts comptaton tme we propose here a hybrd clsterng algorthm called GTFCM that combnes GT FCM wth a fast k-means algorthm for npt data preprocessng of classfcaton algorthms. The proposed algorthm shows mproved comptaton tme when compared wth GT FCM as well as FGTFCM on fve benchmarks from UCI lbrary. eywords- General tye- fzzy, k-means, clsterng, npt data preprocessng, classfcaton I. INTODUCTION Classfcaton s a common problem data mnng [] where datasets are mapped nto predefned grops called classes. Classes are defned accordng to the smlarty of characterstcs or featres of data [1]. Becase, the classes are determned before applyng the real data, ths method s known as a spervsed learnng algorthm. Classfcaton s sed n many felds and scences sch as, bo-nformatcs [15], genetcs [14], bology [16] and healthcare [17]. Several researches have shown that the comptatonal effcency of classfcaton s enhanced f the npt data s frst clstered before for classfcaton. Ths s partclarly applcable when handlng bg data, where low comptaton tme s eqally or more mportant than classfcaton accracy. The class nformaton also mproves the accracy of clsterng [1]. To have the advantages of both clsterng and classfcaton, many exstng algorthms have been developed n a seqental hybrd way [1]. For example n both [11] and [1], frst the crteron s preprocessed and optmzed by a clsterng algorthm and then n the next step the classfcaton crteron s connected wth the acheved clsterng reslts to enhance the accracy of classfcaton algorthms. Generally, there s a trade-off between accracy and comptaton tme of clsterng algorthms,.e. the hgher the accracy, the more the comptaton tme. One of the better known clsterng algorthms s k-means. -means s fast bt has low accracy [1]. On the other hand, general type- fzzy clsterng (GT Alreza owhanmanesh Department of Electrcal Engneerng, Unversty of Neyshabr Neyshabr, Iran rowhanmanesh@eee.org FCM) s a new method that has hgh accracy bt s comptatonally ntensve. In [4], a general type- fzzy clsterng algorthm s ntrodced that s based on α-planes. Ths algorthm has hgh accracy and can deal wth the ncertanty n datasets, whle k-means and FCM, whch are fast clsterng algorthms, cannot handle the ncertanty n a dataset. There are several works that focs on enhancng the speed sse of type- fzzy clsterng. A modfed verson of type- fzzy system was proposed n [6] to mprove the speed (comptatonal tme) of type- fzzy clsterng. Also, n [7]- [9] nterval type- fzzy s sed nstead of general type- fzzy for clsterng, becase nterval type- s faster than general type- fzzy. In addton Yang worked on smlarty measrements of type- fzzy clsterng algorthms on fzzy datasets [19]-[]. In these work, they redefned new smlarty measrements based on non, maxmm. These new smlarty measres affect type- fzzy clsterng performance. In ths paper, we propose a hybrd clsterng algorthm for data npt preprocessng of classfcaton algorthms to address the hgh comptaton tme of general type- fzzy clsterng algorthm. The proposed hybrd method s based on a combnaton of general type- fzzy, whch s an accrate algorthm and k-means, whch s a fast algorthm. We call the proposed approach GTFCM. GTFCM has the advantages of both general type- fzzy and k-means clsterng algorthms,.e. t has hgh accracy and low comptatonal tme. The reslts are compared wth GT FCM and FGTFCM clsterng algorthms for dfferent datasets. Unlabeled datasets are sed for clsterng algorthms; however, labeled datasets are sed for classfcaton algorthms. Snce, we se classfcaton datasets n or experments; we can measre the accracy of or clsterng algorthm. The paper s organzed as follows: secton II dscsses the proposed hybrd algorthm. The reslts and conclson are presented n sectons III and IV, respectvely. II. POPOSED METHOD Or method s based on k-means and general type- fzzy clsterng. General type- fzzy clsterng was presented n [4]. Frst, a general overvew of type- fzzy s gven, and then the proposed method s descrbed /14/$ IEEE 1131
2 A. General Type- Fzzy Clsterng There are two knds of type- fzzy sets whch are often sed n clsterng algorthms: 1) nterval and ) general. In nterval type- fzzy, the secondary membershp fncton s always one, whle n general type- fzzy t s a vale n range of [0,1]. General type- fzzy clsterng s based on FCM (Fzzy C-Means) algorthm. So, the same as FCM, t ntalzes the centers randomly. The FCM algorthm ses lngstc terms sch as Small, Medm or Hgh, modeled by type-1 fzzy sets for the fzzfer parameter M (Fg.1). The FCM algorthm s sed by the GT FCM clster membershp fnctons. The general type- fzzy clsterng proposed n [4] ses α-planes. α-planes manage the ncertanty of general type- fzzy sets. The GT FCM algorthm explots the lngstc fzzfer M for ts secondary membershp fnctons of the general type- fzzy partton matrx as shown n (1). Eqaton (), that s a membershp grade ) expressed as type-1 fzzy sets, s sed to descrbe the membershp degree of pattern x to clster v. ) (1) x X (4): ) [0,1] [0,1] [ s S ), s s ) )] Where s ) and ) are calclated by (3) and s max s ) c l 1 ) d d l 1 sm ( ) 1, c l 1 d d l 1 () sm ( ) (4) mn, c c d sm ( ) 1 s d M ( ) 1 l 1 dl l 1 dl Accordng to [4], centrod C can be calclated as a weghted composton of the nterval centrods of ndvdal α- planes sng (5). The npt of (5) s. Here, d s the dstance of th data from th centrod. Intal centrods are sed for the frst teraton. s and s are obtaned as shown n M M Fg. for each α-planes and c s the nmber of clsters. (3) Fg. 1. ngstc varables for ntalzng the membershp fnctons [4] To compte the precse clster poston, (6) s sed to defzzfy the clster centrod C. [0,1] [ c ( ), c C (5) v 1 1 y C C (6) ( )] In (6), s the nmber of steps that the doman of the centrod has been dscretzed nto and y s the poston vector of th dscretzed step. In ths algorthm the hard-parttonng s done based on the defzzfed vale of the type-1 fzzy membershp grade. So, the followng rle s sed for hardparttonng: If ( ) k )), k 1,..., c, k (7) Then x Clster Bt n [4] formla (8) s sed for hard-parttonng nstead of (7). In (7), snce the Ecldan dstance norm s sed to calclate the membershp of pattern x to clster n the mltdmensonal space, t seems redndant to separately aggregate dentcal membershp vales for each dmenson. So n [4] the athors se (8) for hard-parttonng: ( ) c ( ) ), k 1,..., c, k If ( c x k x Then x Clster (8) The centrod of the type-1 fzzy membershp grade c ) can be calclated sng (9): c ) 1 1 y In ths eqaton, and y have the same defntons as n c ) s the centrod of the th clster. (6), where (9) 113
3 Schematc vew of GT FCM s depcted n Fg.. andom selecton of ntal centrods cases the algorthm to have more teratons, hence, more comptaton tme (lower speed). So, f a clsterng algorthm, sch as k-means, fnds the centers one step before GT FCM and passes them to GT FCM, the comptaton tme of GT FCM wll be redced. α- Plane epresentaton α=0 S M (0), SM (0) (0) α= 1 S M ( 1 ), SM ( 1) ( 1 ) 1 M ( S ) Fzzfer M M ( S ) B. -means -means s one of the most common algorthms n clsterng. In ths method, k denotes the nmber of clsters. -means algorthm has three steps ncldng: Step 1) k clster centers are specfed, randomly.e. one center for each clster, step ) for each npt, dstance from each clster center s calclated. The data belongs to the clster whch has the closest dstance from the center. Ths step s repeated for all npt data, and step 3) the barycenters of clsters (whch are generated n step ) are calclated and consdered as new clster centers and then the algorthm goes to step [1]. These steps are repeated ntl centers do not change for the two consectve teratons. The goal of ths algorthm s to mnmze ts cost fncton denoted as (10) [1],[3]. n J x c (10) 1 1 Here, n s the nmber of samples, s the nmber of clsters, c shows the th clster and x shows th sample of pattern. In ths paper we se Ecldan dstance whch s a tradtonal metrc for dstance measrement of k-means. Followng eqaton presents the Ecldan dstance [10]: n d, y) y ) (11) 1 In ths eqaton d, y) s the Ecldean dstance between sample x wth clster y. C. Prevos work To mprove the comptaton tme of GT FCM, the algorthm at [10] ses the otpt of k-means algorthm as the npt for FCM. Then, the otpts of FCM are sed as the ntal clster centrods for GT FCM. In the followng the detals are dscssed. In the conventonal FCM, the ntal vales of membershp fnctons are random nmbers n the range of [0, 1]. However, f the ntal vales of membershp fnctons are selected more wsely, t s expected to need less nmber of teratons by FCM, hence, the comptaton tme s mproved. To do so, the athors of [10] se k-means to determne the centrods of npt data and then calclate the dstances of each data from all centrods. The normalzed dstances are assmed as ntal vales of membershp fnctons of npt data of FCM. By dong so, FCM wold have a better startng pont and t helps to redce the execton tme and teratons of FCM. Ths algorthm s called FGTFCM. However, sng both k-means and FCM clsterng algorthms for ntalzng GT FCM cases almost a large overhead. X α= S 1 M ( 1), SM ( 1) Type- edcton ( 1) C [ c ( ), c ( )] [01,] α=1 S (1), (1) M S M (1) Fg.. Schematc vew of GT FCM [4] Center Update D. Proposed Method To make the algorthm even faster, we can omt one of the k-means or FCM algorthms whch generate ntal centrods for GT FCM. We omt FCM algorthm from the flow becase k-means s faster than FCM. Also, clsters centrods detected by FCM and k-means clsterng algorthms are close to each other. However, the complexty tme of k-means and FCM are O(ncd) and O(ndc),respectvely [3], where, n denotes nmber of dmensons, s the nmber of teratons, n shows the nmber of sample of dataset and c s the nmber of clsters. Accordng to [3], for n=100, d=3, =0 and for a constant nmber of dataset s samples, the elapsed tme for k- means and FCM are and seconds, respectvely. Also, f we assme the nmber of clsters as a constant and assme n=150, d=, c= and =0, then the tme complexty of k-means and FCM are 1000 and 4000, respectvely [3]. So, FCM needs more comptaton tme than k-means. Hence, n the revsed verson st k-means s employed to fnd the centrods of GT FCM. Ths change mproves the comptaton tme of the algorthm and conseqently, the new algorthm s exected mch faster than FGTFCM. Ths new algorthm s called GTFCM. The flowchart of GTFCM s depcted n Fg.3. As depcted n Fg.3, frst of all the npt dataset are appled to the k-means algorthm. Then, k-means clsters npt dataset and fnds ther centrods. After that, these centrods are sed by GT FCM as the ntal centrods. In the next step, the type- fzzfer fncton calclates the secondary membershp fnctons based on α-planes and sng (3), (4) and Medm lngstc term for secondary membershp fncton as depcted n Fg.1. We se 10 α-planes. Then, EM 1 algorthm [13] s sed for type redcton and fndng the centrods of α-planes. EM ntrodced by Mendel and W to enhance the comptaton tme of M. EM s 39% faster than M algorthm and can save abot two teratons whle M fnd the answer sally between two to sx teratons [13]. In or proposed method, for type redcton we se (5) whch are based on EM algorthm, to fnd the centrods of 10 α-planes. 1 Enhancement arnk Mendel 0 m yc 1 v C ( y )
4 Dong so, the type- fzzy membershp fncton redces to type-1 fzzy whch s a prmary membershp fncton. To fnd the precse center of each clster, the centrods shold be determned sng (6). The centrods calclated by (6) are checked wth prevos centrods of each clster. If they are dfferent, the algorthm recalclates the secondary membershp fncton sng new centrods and then the followng steps are repeated. Otherwse, the algorthm fnshes (Fg. 3). Start Intalze centrods wth k-means clsterng algorthm and se ts otpt as ntal centrods of GT FCM Type- fzzfer and prepare secondary membershp fncton wth 10 α- planes. Fnd No oad Dataset For type redcton compte the centrod of α- plane (10) v as the centrod of Is the new centrod eqal to last centrod? End Yes Classfcaton algorthm th clster. GT FCM Fg. 3. Flowchart of proposed GTFCM TABE I IST OF DATASETS THAT USED FO EXPEIMENTS Attrbtes Dataset Clsters Sze Irs Wne Pma Indans Shttle 43,500 9 Magc 19,00 10 most nmeros data class whch s 80% of data and the second class (class ) contans the remanng less nmeros data classes whch s the 0% of data. All of the datasets of table I are appled to GT FCM, FGTFCM and GTFCM 50 tmes. The machne sed for dong the experments and smlatons s an Acer 5750G system wth an Intel Core 7-630QM@.00GHz and 6.00 GB AM and rnnng Wndows 7. MATAB software has been sed for mplementng the algorthms. For far comparsons of comptaton tme of the three algorthms, the target accracy has been assmed the same for all the three algorthms n all the experments. All of the three algorthms (.e. GT FCM, FGTFCM and GTFCM) are based on GT FCM [4], and se the same membershp fnctons (.e. the same ntal condtons). Snce ntal centrods of k-means and GT FCM are selected randomly, we rn each algorthm for 50 teratons,.e. wth 50 sets of random ntal centrods, to show that the random ntal centrods have trval effects on the reslts. The comptaton tme mprovement (speedp) of GTFCM compared to GT FCM and FGTFCM s calclated sng eqaton (1). Therefore, when the speedp s greater than one the comptaton tme of GTFCM s less than GT FCM or FGTFCM and when the speedp s less than one the comptaton tme of GTFCM s greater than GT FCM or FGTFCM. Comptaton tme mprovement (speedp) = n tme of (GTFCM or FGTFCM) (1) n tme of GTFCM B. Expermental eslts The 30% of Wne dataset whch have been selected randomly s appled to the three algorthms whle the target accracy s assmed to be 66% for all three. The comptaton tme of three algorthms are shown n Fg.4, Fg.5, Fg.10 and Fg.11. III. SIMUATION ESUTS In ths secton the expermental setp and smlaton reslts are presented. A. Expermental Setp In ths paper, fve standard datasets of UCI are selected, ncldng Irs, Wne, Pma Indans, Shttle and Magc whch have been lsted n table I [18]. The Shttle data has been dvded nto two classes. One class (class 1) ncldes the Fg. 4. Comparson of comptaton tme for 50 teratons of three algorthms wth a target accracy of 66% for 53 data of Wne dataset 1134
5 In another experment, 70% of Irs dataset whch have been selected randomly s appled to the three algorthms whle the target accracy s assmed to be 70% for all three algorthms. Comparng table II and table III, reveals that GTFCM otperforms GT FCM and FGTFCM sgnfcantly for low target accraces. For the experments performed for generatng reslts of table II and III, 70% of each dataset whch selected randomly, were sed. However, for the experments done for generatng reslts of Fg.4 and 5, 30% of each dataset, whch selected randomly, were exploted. Snce, n ths paper we focs on comptaton tme redcton and not accracy, the target accraces are selected close to the maxmm accracy. TABE II COMPAING COMPUTATION TIME (IN SECONDS) OF THEE AGOITHMS WITH OW TAGET ACCUACY Method GTFCM (n seconds) FGTFCM (n seconds) GTFCM (n seconds) Speedp of GT- FCM vs. GT FCM Speedp of GT- FCM vs. FGTFCM a Target accracy Irs Acc a : 50% Wne 50% Pma Indans 65% Shttle 70% e-5.3e e-5 1.7e-5 1.7e-5.37e-4.51e-5 1.5e e-5.31e As llstrated n Fg.4 and Fg.5, the comptaton tme of FGTFCM and GTFCM enhanced sgnfcantly compared to GT FCM. The average speedp of GTFCM compared to GT FCM and FGTFCM are and respectvely for Fg. 4 and are 1509 and 1.04, for Fg. 5, correspondngly. Becase the comptaton tme of GT FCM s mch larger than the two others, the dfferences of FGTFCM and GTFCM comptaton tme s not seen n these fgre. Therefore, two other Fg. 6 and 7, whch are zoomed n of Fg. 4 and 5, have been added for comparng only FGTFCM aganst GTFCM. The same as Fg.4 and 5, n Fg. 6 and 7 the reslts of FGTFCM and GTFCM are very smlar. Therefore to clarfy the dfference of these clsterng algorthms, n Fg. 7 and Fg.8, the same reslts are redrawn st for 35 teratons. TABE III COMPAING COMPUTATION TIME OF THEE AGOITHMS (IN SECONDS) ON DIFFEENT DATASETS WITH DIFFEENT TAGET ACCUACY Method GTFCM (n seconds) FGTFCM (n seconds) GTFCM (n seconds) Speedp of GT- FCM vs. GT FCM Speedp of GT- FCM vs. FGTFCM Irs Acc a : 86% Wne 71% Pma Indans 68% Shttle 74.7% e e e e e-5.30e e a Target accracy Fg. 5. Comparson of comptaton tme for 50 teratons of three algorthms wth a target accracy of 70% for 105 data of Irs dataset Fg. 6. Comparson of comptaton tme for 50 teratons of two algorthms wth a target accracy of 70% for 105 data of Irs dataset 1135
6 Fg. 7. Comparson of comptaton tme between two algorthms to reach to the 66% accracy n 50 tmes teratons for 53 data of Wne dataset Fg. 9. Comparson of comptaton tme between two algorthms to reach to the 66% accracy n 35 tmes teratons for 53 data of Wne dataset Fg. 8. Comparson of comptaton tme for 35 teratons of two algorthms wth a target accracy of 70% for 105 data of Irs dataset Fg. 10 shows the comptaton tme for GT FCM, FGTFCM and GTFCM, when the accracy s 60% and 30% of Pma Indans dataset s selected. The average comptaton tme mprovement of GTFCM compared to GT FCM and FGTFCM for ths dataset are 344 and 1911, respectvely. The same experments have been done sng 70% of Irs dataset assmng 75% target accracy. The reslts have been depcted n Fg. 11. The average comptaton tme mprovement of GTFCM compared to GT FCM and FGTFCM for ths dataset are and 0.55, respectvely. The reason for performance degradaton of GTFCM compared to FGTFCM s de to the last fve rns of FGTFCM whch have long rntme. Fg. 10. Comparson of comptaton tme for 50 teratons of three algorthms wth a target accracy of 60% for 30 data of Pma Indans dataset De to the large dfferences between execton tme of the last 15 rns of FGTFCM and the other 35 rns, t seems that crves of FGTFCM and GTFCM are very close to each other for the frst 35 rns. However, the real dfferences of FGTFCM and GTFCM comptaton tme are not llstrated n Fg. 11, clearly. To clarfy ths, Fg. 1, whch s zoomed n of Fg. 11, has been added for comparng FGTFCM and GTFCM for the frst 35 rns. Consderng only the frst 35 rns, the average execton tme mprovement of GTFCM compared to FGTFCM for ths dataset s The averages of comptaton tme (n seconds) of 50 teratons of each algorthm on a specfc dataset wth specfc target accracy are avalable n table II and table III. Also, table II and III show the speedp of GTFCM compared to FGTFCM and GT FCM for each specfc dataset wth specfc target accracy. The npt data sed for these tables are 70% of data of each dataset that were selected randomly. 1136
7 Accordng to table II the maxmm speedp for GTFCM vs. FGTFCM s and for GTFCM vs. GT FCM s and for table III are and 306, respectvely. Totally, we tred all of these datasets for fve dfferent accraces. Bt, de to page lmtaton we only presented two grops of the reslts n tables II and III and the rest of the reslts were not presented. The average speedp for all of the reslts obtaned for GTFCM vs. FGTFCM and GT FCM, are 1818 and 7456, respectvely. Accordng to table II and table III, FGTFCM obtans better reslts for Irs dataset. The reason s that the combnaton of k-means and FCM prodces better ntal centrods compared to k-means, whch shortens GT FCM execton tme. However, the combnaton of k-means and FCM prodces worse ntal centrods compared to the proposed method for Wne dataset whch lengthens GT FCM execton tme. The proposed method obtans better reslts for 80% of case stdes sed for tables II and III. The best reslts are bolded n table II and table III. Fg. 11. Comparson of comptaton tme for 50 teratons of three algorthms wth a target accracy of 75% for 105 data of Irs dataset Fg. 1. Comparson of comptaton tme for 35 teratons of two algorthms wth a target accracy of 75 % for 105 data of Irs dataset In table III, for the Shttle dataset, the GT FCM performs better compared to FGTFCM and GTFCM. The reason s that GT FCM reaches to the target accracy, n the frst teraton. So, n both FGTFCM and GTFCM algorthms, k-means and FCM algorthms whch are sed to ntalze the centrod of GT FCM, reslts n longer comptaton tme for FGTFCM and GTFCM. Therefore, these algorthms rn slower than GT FCM. Also, n table III, for Irs dataset, FGTFCM s faster than GTFCM. The reason of ths phenomenon s that for teratons from 50 teratons of GTFCM are very tme consmng. The effects of these slow teratons cases the average tme of GTFCM algorthm to be lower than FGTFCM. Consderng k- means, n whch ts ntal centrods are selected randomly, reveals that f the random centrods are not selected properly, t reslts n poor accracy and therefore cases GT FCM to need more teraton and hence, makes GTFCM slower. The same problem happens n fgres 6 and 7. Accordng to table II and III, althogh, n several cases GTFCM s slower than GT FCM and FGTFCM, the comptaton tme dfferences of GTFCM n these cases s very small and are approxmately 4.0e-7 seconds compared to GT FCM and FGTFCM. IV. CONCUSION ecently, several works have sed clsterng and classfcaton n seqental strctres to mprove the performance of classfcaton algorthms. As they ndcated, the effcency of classfcaton learnng s enhanced f the npt data s frst clstered and then sed for classfcaton. However, there s a trade-off between comptaton tme and accracy of clsterng algorthms. In ths paper, a new clsterng method s ntrodced to mprove the comptaton tme of a classfcaton algorthm by preprocessng classfcaton dataset. To address the conflct of hgh comptaton tme and hgh accracy of clsterng algorthm, we propose a hybrd clsterng algorthm called GTFCM. Ths algorthm s a combnaton of hgh accracy general type- fzzy C-means (GT FCM) that can deal wth ncertanty va sng α-planes wth low comptaton tme k- means algorthm for npt data preprocessng of classfcaton algorthms. The proposed algorthm mproves the speed of GT FCM and rn on fve datasets of UCI for clsterng wth dfferent target accracy. GTFCM has better effcency compared to GT FCM for almost all cases. The reason s that GTFCM prodces better ntal centrods for GT FCM. The average speedp of GTFCM compared to GT FCM on fve datasets ncldng Irs, Wne, Pma Indans, Shttle and Magc, s In 86% of case stdes; GTFCM s faster than FGTFCM. The reason s that for these cases combnaton of k-means and FCM takes long tme to converge to good ntal centrods for GT FCM, however GTFCM does not have the overhead of FCM. In the remanng 14% cases, becase the combnaton of k-means and FCM prodces better ntal centrods for GT FCM, FGTFCM has better speedp compared to GTFCM. Fnally, the average speedp of GTFCM compared to FGTFCM on fve datasets ncldng Irs, Wne, Pma Indans, Shttle and Magc, s In conclson, the proposed method 1137
8 (GTFCM) s faster than GT FCM and FGTFCM by 7456 and 1818, respectvely. ACNOWEDGMENT We thank Dr. Ondre nda from Idaho Unversty, for hs generosty and spportng s. EFEENCES [1]. Fnats and. Hasegawa, New fndamental technologes n data mnng. Frst pblshed Janary, 011.Prnted n Inda. ISBN []. Athada, M. Tssera, C. Fernando. Data Mnng Applcatons: Promse and Challenges. Data Mnng and nowledge Dscovery n eal fe Applcatons, ISBN , pp. 438, Febrary 009, I-Tech, Venna, Astra. [3]. X, D. Wnsch II, Srvey of Clsterng Algorthms. IEEE Transactons On Neral Networks, Vol. 16, No. 3, MAY 005. [4] O. nda, M. Manc, General Type- Fzzy C-Means Algorthm for Uncertan Fzzy Clsterng. Fzzy Systems, IEEE Transactons. Feb 13 01, ISSN : [5] A.. Jan, Data clsterng: 50 years beyond k-means. Pattern ecognton etters 31 (010) [6] M. H. Fazel Zarand, I. B. Trksen, O. Torab asb, Type- fzzy modelng for deslphrzaton of steel process. Expert Systems wth Applcatons 3 (007) [7] A. Shah,. Bnt Atan and M-D. Nasr Slaman, An effectve fzzy c-mean and type- fzzy logc for weather forecastng. Jornal of Theoretcal and Appled Informaton Technology JATIT. Malaysa. [8] Q. ang and J. Mendel, Decson Feedback Eqalzer for Nonlnear Tme-Varyng Channels Usng Type- fzzy Adaptve Flters. Fzzy Systems, 000. FUZZ IEEE 000. [9] G. Zheng; Jng Wang; Wengwe Zho; Yong Zhang, A Smlarty Measre between Interval Type- Fzzy Sets. Proceedngs of the 010 IEEE, Internatonal Conference on Mechatroncs and Atomaton, Agst 4-7, 010, X'an, Chna. Internatonal Conference on 6-8 Jly 011. [10] V. Nor, Mohammad-. Akbarzadeh-T., Alreza owhanmanesh. General type- fzzy clsterng sng hybrd of k- means and type-1 fzzy clsterng for data preprocessng n Persan. 8 th Symposm on Advances n Scence and Technology. Dec 19, 013. [11] E.. ;Pfahrnger, B. ; Holmes, G. Clsterng for classfcaton. Informaton Technology n Asa (CITA 11), 011 7th Internatonal Conference on Dgtal Obect Identfer: /CITA Pblcaton Year: 011, Page(s): 1 8. [1] W Ca ;S. Chen ;D. Zhang. A Mlt-obectve Smltaneos earnng Framework for Clsterng and Classfcaton. IEEE Transactons on Neral Networks, Volme: 1, Isse:.Pblshed n 010. [13] D. W, J. Mendel. Enhanced arnk-mendel Algorthms for Interval Type- Fzzy Sets and Systems. Fzzy Informaton Processng Socety, 007. NAFIPS '07. Annal Meetng of the North Amercan. [14] Yvara, N.; Vvekanandan, P.; An effcent SVM based tmor classfcaton wth symmetry Non-negatve Matrx Factorzaton sng gene expresson data. Internatonal Conference on Informaton Commncaton and Embedded Systems (ICICES), Feb [15] Borges, H.B. ; Nevola, J.C. Herarchcal classfcaton sng a Compettve Neral Network. Eghth Internatonal Conference on Natral Comptaton (ICNC), May 01. [16] W. Yang ;. Wang ; W. Zo. Predcton of proten secondary strctre sng large margn nearest neghbor classfcaton. Advanced Compter Control (ICACC), 011 3rd Internatonal Conference Jan [17] Swangnetr, M.; aber, D.B.; Emotonal State Classfcaton n Patent obot Interacton Usng Wavelet Analyss and Statstcs-Based Featre Selecton. IEEE Transactons on Hman-Machne Systems (Volme: 43, Isse: 1). Jan [18] A. Frank, A. Asncon, UCI Machne earnng epostory [ Irvne, CA: Unversty of Calforna, School of Informatcs and Compter Scence. [19] Der-Chen n, Mn-Shen Yang, A smlarty measre between type- fzzy sets wth ts applcaton to clsterng. Forth Internatonal Conference on Fzzy Systems and nowledge Dscovery, 007. FSD 007. [0] Wen-lang Hng, Mn-shen Yang, Smlarty Measres Between Type- Fzzy Sets. Internatonal Jornal of Uncertanty, Fzzness and nowledge-based Systems. Vol. 1, No. 6 (004) World Scentfc Pblshng Company. [1] Mn-Shen Yang, Der-Chen n. On smlarty and nclson measres between type- fzzy sets wth an applcaton to clsterng. Compters and Mathematcs wth Applcatons 57 (009) 896_907. [] Hwang C.-M., Yang M.-S., Hng W.-., On smlarty, nclson measre and entropy between type- fzzy sets. Internatonal Jornal of Uncertanty, Fzzness and nowlege-based Systems 01. [3] Som Ghosh, Sanay mar Dbey. Comparatve Analyss of - Means and Fzzy CMeans Algorthms. ((IJACSA) Internatonal Jornal of Advanced Compter Scence and Applcatons, Vol. 4, No.4,
Modeling Local Uncertainty accounting for Uncertainty in the Data
Modelng Local Uncertanty accontng for Uncertanty n the Data Olena Babak and Clayton V Detsch Consder the problem of estmaton at an nsampled locaton sng srrondng samples The standard approach to ths problem
More informationNumerical Solution of Deformation Equations. in Homotopy Analysis Method
Appled Mathematcal Scences, Vol. 6, 2012, no. 8, 357 367 Nmercal Solton of Deformaton Eqatons n Homotopy Analyss Method J. Izadan and M. MohammadzadeAttar Department of Mathematcs, Faclty of Scences, Mashhad
More informationHybrid Method of Biomedical Image Segmentation
Hybrd Method of Bomedcal Image Segmentaton Mng Hng Hng Department of Electrcal Engneerng and Compter Scence, Case Western Reserve Unversty, Cleveland, OH, Emal: mxh8@case.ed Abstract In ths paper we present
More informationParallelism for Nested Loops with Non-uniform and Flow Dependences
Parallelsm for Nested Loops wth Non-unform and Flow Dependences Sam-Jn Jeong Dept. of Informaton & Communcaton Engneerng, Cheonan Unversty, 5, Anseo-dong, Cheonan, Chungnam, 330-80, Korea. seong@cheonan.ac.kr
More informationCluster Analysis of Electrical Behavior
Journal of Computer and Communcatons, 205, 3, 88-93 Publshed Onlne May 205 n ScRes. http://www.scrp.org/ournal/cc http://dx.do.org/0.4236/cc.205.350 Cluster Analyss of Electrcal Behavor Ln Lu Ln Lu, School
More informationAn Improved Isogeometric Analysis Using the Lagrange Multiplier Method
An Improved Isogeometrc Analyss Usng the Lagrange Mltpler Method N. Valzadeh 1, S. Sh. Ghorash 2, S. Mohammad 3, S. Shojaee 1, H. Ghasemzadeh 2 1 Department of Cvl Engneerng, Unversty of Kerman, Kerman,
More informationClassifier Selection Based on Data Complexity Measures *
Classfer Selecton Based on Data Complexty Measures * Edth Hernández-Reyes, J.A. Carrasco-Ochoa, and J.Fco. Martínez-Trndad Natonal Insttute for Astrophyscs, Optcs and Electroncs, Lus Enrque Erro No.1 Sta.
More informationLearning the Kernel Parameters in Kernel Minimum Distance Classifier
Learnng the Kernel Parameters n Kernel Mnmum Dstance Classfer Daoqang Zhang 1,, Songcan Chen and Zh-Hua Zhou 1* 1 Natonal Laboratory for Novel Software Technology Nanjng Unversty, Nanjng 193, Chna Department
More informationSubspace clustering. Clustering. Fundamental to all clustering techniques is the choice of distance measure between data points;
Subspace clusterng Clusterng Fundamental to all clusterng technques s the choce of dstance measure between data ponts; D q ( ) ( ) 2 x x = x x, j k = 1 k jk Squared Eucldean dstance Assumpton: All features
More informationOBJECT TRACKING BY ADAPTIVE MEAN SHIFT WITH KERNEL BASED CENTROID METHOD
ISSN : 0973-739 Vol. 3, No., Janary-Jne 202, pp. 39-42 OBJECT TRACKING BY ADAPTIVE MEAN SHIFT WITH KERNEL BASED CENTROID METHOD Rahl Mshra, Mahesh K. Chohan 2, and Dhraj Ntnawwre 3,2,3 Department of Electroncs,
More informationContent Based Image Retrieval Using 2-D Discrete Wavelet with Texture Feature with Different Classifiers
IOSR Journal of Electroncs and Communcaton Engneerng (IOSR-JECE) e-issn: 78-834,p- ISSN: 78-8735.Volume 9, Issue, Ver. IV (Mar - Apr. 04), PP 0-07 Content Based Image Retreval Usng -D Dscrete Wavelet wth
More informationType-2 Fuzzy Non-uniform Rational B-spline Model with Type-2 Fuzzy Data
Malaysan Journal of Mathematcal Scences 11(S) Aprl : 35 46 (2017) Specal Issue: The 2nd Internatonal Conference and Workshop on Mathematcal Analyss (ICWOMA 2016) MALAYSIAN JOURNAL OF MATHEMATICAL SCIENCES
More informationImprovement of Spatial Resolution Using BlockMatching Based Motion Estimation and Frame. Integration
Improvement of Spatal Resoluton Usng BlockMatchng Based Moton Estmaton and Frame Integraton Danya Suga and Takayuk Hamamoto Graduate School of Engneerng, Tokyo Unversty of Scence, 6-3-1, Nuku, Katsuska-ku,
More informationMachine Learning. Topic 6: Clustering
Machne Learnng Topc 6: lusterng lusterng Groupng data nto (hopefully useful) sets. Thngs on the left Thngs on the rght Applcatons of lusterng Hypothess Generaton lusters mght suggest natural groups. Hypothess
More informationTerm Weighting Classification System Using the Chi-square Statistic for the Classification Subtask at NTCIR-6 Patent Retrieval Task
Proceedngs of NTCIR-6 Workshop Meetng, May 15-18, 2007, Tokyo, Japan Term Weghtng Classfcaton System Usng the Ch-square Statstc for the Classfcaton Subtask at NTCIR-6 Patent Retreval Task Kotaro Hashmoto
More informationNUMERICAL SOLVING OPTIMAL CONTROL PROBLEMS BY THE METHOD OF VARIATIONS
ARPN Journal of Engneerng and Appled Scences 006-017 Asan Research Publshng Network (ARPN). All rghts reserved. NUMERICAL SOLVING OPTIMAL CONTROL PROBLEMS BY THE METHOD OF VARIATIONS Igor Grgoryev, Svetlana
More informationNURBS curve method Taylor's launch type of interpolation arithmetic Wan-Jun Zhang1,2,3,a, Shan-Ping Gao1,b Xi-Yan Cheng 1,c &Feng Zhang2,d
Advances n Engneerng Research volme 118 nd Internatonal Conference on Atomaton Mechancal Control and Comptatonal Engneerng (AMCCE 17) NURBS crve method Taylor's lanch type of nterpolaton arthmetc Wan-Jn
More informationA New Approach For the Ranking of Fuzzy Sets With Different Heights
New pproach For the ankng of Fuzzy Sets Wth Dfferent Heghts Pushpnder Sngh School of Mathematcs Computer pplcatons Thapar Unversty, Patala-7 00 Inda pushpndersnl@gmalcom STCT ankng of fuzzy sets plays
More informationDetermining the Optimal Bandwidth Based on Multi-criterion Fusion
Proceedngs of 01 4th Internatonal Conference on Machne Learnng and Computng IPCSIT vol. 5 (01) (01) IACSIT Press, Sngapore Determnng the Optmal Bandwdth Based on Mult-crteron Fuson Ha-L Lang 1+, Xan-Mn
More informationThe Research of Support Vector Machine in Agricultural Data Classification
The Research of Support Vector Machne n Agrcultural Data Classfcaton Le Sh, Qguo Duan, Xnmng Ma, Me Weng College of Informaton and Management Scence, HeNan Agrcultural Unversty, Zhengzhou 45000 Chna Zhengzhou
More informationHierarchical clustering for gene expression data analysis
Herarchcal clusterng for gene expresson data analyss Gorgo Valentn e-mal: valentn@ds.unm.t Clusterng of Mcroarray Data. Clusterng of gene expresson profles (rows) => dscovery of co-regulated and functonally
More informationA General Algorithm for Computing Distance Transforms in Linear Time
Ths chapter has been pblshed as: A. Mejster, J. B. T. M. Roerdnk and W. H. Hesselnk, A general algorthm for comptng dstance transforms n lnear tme. In: Mathematcal Morphology and ts Applcatons to Image
More informationA Fast Content-Based Multimedia Retrieval Technique Using Compressed Data
A Fast Content-Based Multmeda Retreval Technque Usng Compressed Data Borko Furht and Pornvt Saksobhavvat NSF Multmeda Laboratory Florda Atlantc Unversty, Boca Raton, Florda 3343 ABSTRACT In ths paper,
More informationScheduling with Integer Time Budgeting for Low-Power Optimization
Schedlng wth Integer Tme Bdgetng for Low-Power Optmzaton We Jang, Zhr Zhang, Modrag Potkonjak and Jason Cong Compter Scence Department Unversty of Calforna, Los Angeles Spported by NSF, SRC. Otlne Introdcton
More informationA fast algorithm for color image segmentation
Unersty of Wollongong Research Onlne Faculty of Informatcs - Papers (Arche) Faculty of Engneerng and Informaton Scences 006 A fast algorthm for color mage segmentaton L. Dong Unersty of Wollongong, lju@uow.edu.au
More informationAn Internal Clustering Validation Index for Boolean Data
BULGARIAN ACADEMY OF SCIENCES CYBERNETICS AND INFORMATION TECHNOLOGIES Volume 16, No 6 Specal ssue wth selecton of extended papers from 6th Internatonal Conference on Logstc, Informatcs and Servce Scence
More informationIncremental Learning with Support Vector Machines and Fuzzy Set Theory
The 25th Workshop on Combnatoral Mathematcs and Computaton Theory Incremental Learnng wth Support Vector Machnes and Fuzzy Set Theory Yu-Mng Chuang 1 and Cha-Hwa Ln 2* 1 Department of Computer Scence and
More informationTsinghua University at TAC 2009: Summarizing Multi-documents by Information Distance
Tsnghua Unversty at TAC 2009: Summarzng Mult-documents by Informaton Dstance Chong Long, Mnle Huang, Xaoyan Zhu State Key Laboratory of Intellgent Technology and Systems, Tsnghua Natonal Laboratory for
More informationA Deflected Grid-based Algorithm for Clustering Analysis
A Deflected Grd-based Algorthm for Clusterng Analyss NANCY P. LIN, CHUNG-I CHANG, HAO-EN CHUEH, HUNG-JEN CHEN, WEI-HUA HAO Department of Computer Scence and Informaton Engneerng Tamkang Unversty 5 Yng-chuan
More informationEasyChair Preprint. Laplacian Deep Hashing for Image Retrieval
EasyChar Preprnt 222 Laplacan Deep Hashng for Image Retreval Chnzh Wang, Fangy Zho, Lngy Yan, Zhwe Ye, Pan W and Hanln L EasyChar preprnts are ntended for rapd dssemnaton of research reslts and are ntegrated
More informationMULTISPECTRAL IMAGES CLASSIFICATION BASED ON KLT AND ATR AUTOMATIC TARGET RECOGNITION
MULTISPECTRAL IMAGES CLASSIFICATION BASED ON KLT AND ATR AUTOMATIC TARGET RECOGNITION Paulo Quntlano 1 & Antono Santa-Rosa 1 Federal Polce Department, Brasla, Brazl. E-mals: quntlano.pqs@dpf.gov.br and
More informationAn Application of the Dulmage-Mendelsohn Decomposition to Sparse Null Space Bases of Full Row Rank Matrices
Internatonal Mathematcal Forum, Vol 7, 2012, no 52, 2549-2554 An Applcaton of the Dulmage-Mendelsohn Decomposton to Sparse Null Space Bases of Full Row Rank Matrces Mostafa Khorramzadeh Department of Mathematcal
More informationAn Optimal Algorithm for Prufer Codes *
J. Software Engneerng & Applcatons, 2009, 2: 111-115 do:10.4236/jsea.2009.22016 Publshed Onlne July 2009 (www.scrp.org/journal/jsea) An Optmal Algorthm for Prufer Codes * Xaodong Wang 1, 2, Le Wang 3,
More informationClustering Algorithm Combining CPSO with K-Means Chunqin Gu 1, a, Qian Tao 2, b
Internatonal Conference on Advances n Mechancal Engneerng and Industral Informatcs (AMEII 05) Clusterng Algorthm Combnng CPSO wth K-Means Chunqn Gu, a, Qan Tao, b Department of Informaton Scence, Zhongka
More informationFrom Comparing Clusterings to Combining Clusterings
Proceedngs of the Twenty-Thrd AAAI Conference on Artfcal Intellgence (008 From Comparng Clusterngs to Combnng Clusterngs Zhwu Lu and Yuxn Peng and Janguo Xao Insttute of Computer Scence and Technology,
More informationUsing Fuzzy Logic to Enhance the Large Size Remote Sensing Images
Internatonal Journal of Informaton and Electroncs Engneerng Vol. 5 No. 6 November 015 Usng Fuzzy Logc to Enhance the Large Sze Remote Sensng Images Trung Nguyen Tu Huy Ngo Hoang and Thoa Vu Van Abstract
More informationA Comparative Study of Constraint-Handling Techniques in Evolutionary Constrained Multiobjective Optimization
A omparatve Stdy of onstrant-handlng Technqes n Evoltonary onstraned Mltobectve Optmzaton Ja-Peng L, Yong Wang, Member, IEEE, Shengxang Yang, Senor Member, IEEE, and Zxng a, Senor Member, IEEE Abstract
More informationBIN XIA et al: AN IMPROVED K-MEANS ALGORITHM BASED ON CLOUD PLATFORM FOR DATA MINING
An Improved K-means Algorthm based on Cloud Platform for Data Mnng Bn Xa *, Yan Lu 2. School of nformaton and management scence, Henan Agrcultural Unversty, Zhengzhou, Henan 450002, P.R. Chna 2. College
More informationA Binarization Algorithm specialized on Document Images and Photos
A Bnarzaton Algorthm specalzed on Document mages and Photos Ergna Kavalleratou Dept. of nformaton and Communcaton Systems Engneerng Unversty of the Aegean kavalleratou@aegean.gr Abstract n ths paper, a
More informationNetwork Intrusion Detection Based on PSO-SVM
TELKOMNIKA Indonesan Journal of Electrcal Engneerng Vol.1, No., February 014, pp. 150 ~ 1508 DOI: http://dx.do.org/10.11591/telkomnka.v1.386 150 Network Intruson Detecton Based on PSO-SVM Changsheng Xang*
More informationObstacle Avoidance by Using Modified Hopfield Neural Network
bstacle Avodance by Usng Modfed Hopfeld Neral Network Panrasee Rtthpravat Center of peraton for Feld Robotcs Development (FIB), Kng Mongkt s Unversty of Technology Thonbr. 91 Sksawas road Tongkr Bangkok
More informationRecommended Items Rating Prediction based on RBF Neural Network Optimized by PSO Algorithm
Recommended Items Ratng Predcton based on RBF Neural Network Optmzed by PSO Algorthm Chengfang Tan, Cayn Wang, Yuln L and Xx Q Abstract In order to mtgate the data sparsty and cold-start problems of recommendaton
More informationA new selection strategy for selective cluster ensemble based on Diversity and Independency
A new selecton strategy for selectve cluster ensemble based on Dversty and Independency Muhammad Yousefnezhad a, Al Rehanan b, Daoqang Zhang a and Behrouz Mnae-Bdgol c a Department of Computer Scence,
More informationAn Image Fusion Approach Based on Segmentation Region
Rong Wang, L-Qun Gao, Shu Yang, Yu-Hua Cha, and Yan-Chun Lu An Image Fuson Approach Based On Segmentaton Regon An Image Fuson Approach Based on Segmentaton Regon Rong Wang, L-Qun Gao, Shu Yang 3, Yu-Hua
More informationProblem Definitions and Evaluation Criteria for Computational Expensive Optimization
Problem efntons and Evaluaton Crtera for Computatonal Expensve Optmzaton B. Lu 1, Q. Chen and Q. Zhang 3, J. J. Lang 4, P. N. Suganthan, B. Y. Qu 6 1 epartment of Computng, Glyndwr Unversty, UK Faclty
More informationSupport Vector Machines
/9/207 MIST.6060 Busness Intellgence and Data Mnng What are Support Vector Machnes? Support Vector Machnes Support Vector Machnes (SVMs) are supervsed learnng technques that analyze data and recognze patterns.
More informationThe Greedy Method. Outline and Reading. Change Money Problem. Greedy Algorithms. Applications of the Greedy Strategy. The Greedy Method Technique
//00 :0 AM Outlne and Readng The Greedy Method The Greedy Method Technque (secton.) Fractonal Knapsack Problem (secton..) Task Schedulng (secton..) Mnmum Spannng Trees (secton.) Change Money Problem Greedy
More informationPerson Identity Clustering in TV Show Videos
Person Identty Clsterng n TV Show Vdeos Yna Han*, Gzhong L* *School of Electrcal and Informaton Engneerng, X an Jaotong Unversty, X an, P.R.Chna Emal:yan@malst.xjt.ed.cn, lgz@mal.xjt.ed.cn Keywords: Identty
More informationCorner-Based Image Alignment using Pyramid Structure with Gradient Vector Similarity
Journal of Sgnal and Informaton Processng, 013, 4, 114-119 do:10.436/jsp.013.43b00 Publshed Onlne August 013 (http://www.scrp.org/journal/jsp) Corner-Based Image Algnment usng Pyramd Structure wth Gradent
More informationA Fast Visual Tracking Algorithm Based on Circle Pixels Matching
A Fast Vsual Trackng Algorthm Based on Crcle Pxels Matchng Zhqang Hou hou_zhq@sohu.com Chongzhao Han czhan@mal.xjtu.edu.cn Ln Zheng Abstract: A fast vsual trackng algorthm based on crcle pxels matchng
More informationBRDPHHC: A Balance RDF Data Partitioning Algorithm based on Hybrid Hierarchical Clustering
015 IEEE 17th Internatonal Conference on Hgh Performance Computng and Communcatons (HPCC), 015 IEEE 7th Internatonal Symposum on Cyberspace Safety and Securty (CSS), and 015 IEEE 1th Internatonal Conf
More informationTHE CONDENSED FUZZY K-NEAREST NEIGHBOR RULE BASED ON SAMPLE FUZZY ENTROPY
Proceedngs of the 20 Internatonal Conference on Machne Learnng and Cybernetcs, Guln, 0-3 July, 20 THE CONDENSED FUZZY K-NEAREST NEIGHBOR RULE BASED ON SAMPLE FUZZY ENTROPY JUN-HAI ZHAI, NA LI, MENG-YAO
More informationA Clustering Algorithm for Key Frame Extraction Based on Density Peak
Journal of Computer and Communcatons, 2018, 6, 118-128 http://www.scrp.org/ournal/cc ISSN Onlne: 2327-5227 ISSN Prnt: 2327-5219 A Clusterng Algorthm for Key Frame Extracton Based on Densty Peak Hong Zhao
More informationConcurrent Apriori Data Mining Algorithms
Concurrent Apror Data Mnng Algorthms Vassl Halatchev Department of Electrcal Engneerng and Computer Scence York Unversty, Toronto October 8, 2015 Outlne Why t s mportant Introducton to Assocaton Rule Mnng
More informationFactorization vs. Regularization: Fusing Heterogeneous Social Relationships in Top-N Recommendation
Factorzaton vs. Reglarzaton: Fsng Heterogeneos Socal Relatonshps n Top-N Recommendaton Qan Yan IBM Research - Chna Zhonggancn Software Park, Hadan Dstrct Bejng, Chna qanyan@cn.bm.com L Chen Department
More informationVirtual Machine Migration based on Trust Measurement of Computer Node
Appled Mechancs and Materals Onlne: 2014-04-04 ISSN: 1662-7482, Vols. 536-537, pp 678-682 do:10.4028/www.scentfc.net/amm.536-537.678 2014 Trans Tech Publcatons, Swtzerland Vrtual Machne Mgraton based on
More informationMachine Learning: Algorithms and Applications
14/05/1 Machne Learnng: Algorthms and Applcatons Florano Zn Free Unversty of Bozen-Bolzano Faculty of Computer Scence Academc Year 011-01 Lecture 10: 14 May 01 Unsupervsed Learnng cont Sldes courtesy of
More informationK-means Optimization Clustering Algorithm Based on Hybrid PSO/GA Optimization and CS validity index
Orgnal Artcle Prnt ISSN: 3-6379 Onlne ISSN: 3-595X DOI: 0.7354/jss/07/33 K-means Optmzaton Clusterng Algorthm Based on Hybrd PSO/GA Optmzaton and CS valdty ndex K Jahanbn *, F Rahmanan, H Rezae 3, Y Farhang
More informationPerformance Modeling of Web-based Software Systems with Subspace Identification
Acta Poltechnca Hngarca Vol. 13, No. 7, 2016 Performance Modelng of Web-based Software Sstems wth Sbspace Identfcaton Ágnes Bogárd-Mészöl, András Rövd, Shohe Yokoama Department of Atomaton and Appled Informatcs,
More informationGA-Based Learning Algorithms to Identify Fuzzy Rules for Fuzzy Neural Networks
Seventh Internatonal Conference on Intellgent Systems Desgn and Applcatons GA-Based Learnng Algorthms to Identfy Fuzzy Rules for Fuzzy Neural Networks K Almejall, K Dahal, Member IEEE, and A Hossan, Member
More informationFingerprint matching based on weighting method and SVM
Fngerprnt matchng based on weghtng method and SVM Ja Ja, Lanhong Ca, Pnyan Lu, Xuhu Lu Key Laboratory of Pervasve Computng (Tsnghua Unversty), Mnstry of Educaton Bejng 100084, P.R.Chna {jaja}@mals.tsnghua.edu.cn
More informationRobust Visual Tracking via Fuzzy Kernel Representation
Research Jornal of Appled Scences, Engneerng and Technolog 5(): 3-38, 3 ISSN: 4-7459; e-issn: 4-7467 Maxwell Scentfc Organzaton, 3 Sbmtted: October 7, Accepted: December, Pblshed: Aprl 5, 3 Robst Vsal
More informationA Combined Approach for Mining Fuzzy Frequent Itemset
A Combned Approach for Mnng Fuzzy Frequent Itemset R. Prabamaneswar Department of Computer Scence Govndammal Adtanar College for Women Truchendur 628 215 ABSTRACT Frequent Itemset Mnng s an mportant approach
More informationEVALUATION OF THE PERFORMANCES OF ARTIFICIAL BEE COLONY AND INVASIVE WEED OPTIMIZATION ALGORITHMS ON THE MODIFIED BENCHMARK FUNCTIONS
Academc Research Internatonal ISS-L: 3-9553, ISS: 3-9944 Vol., o. 3, May 0 EVALUATIO OF THE PERFORMACES OF ARTIFICIAL BEE COLOY AD IVASIVE WEED OPTIMIZATIO ALGORITHMS O THE MODIFIED BECHMARK FUCTIOS Dlay
More informationDocument Representation and Clustering with WordNet Based Similarity Rough Set Model
IJCSI Internatonal Journal of Computer Scence Issues, Vol. 8, Issue 5, No 3, September 20 ISSN (Onlne): 694-084 www.ijcsi.org Document Representaton and Clusterng wth WordNet Based Smlarty Rough Set Model
More informationModule Management Tool in Software Development Organizations
Journal of Computer Scence (5): 8-, 7 ISSN 59-66 7 Scence Publcatons Management Tool n Software Development Organzatons Ahmad A. Al-Rababah and Mohammad A. Al-Rababah Faculty of IT, Al-Ahlyyah Amman Unversty,
More informationX- Chart Using ANOM Approach
ISSN 1684-8403 Journal of Statstcs Volume 17, 010, pp. 3-3 Abstract X- Chart Usng ANOM Approach Gullapall Chakravarth 1 and Chaluvad Venkateswara Rao Control lmts for ndvdual measurements (X) chart are
More informationRestaurants Review Star Prediction for Yelp Dataset
Restarants Revew Star Predcton for Yelp Dataset Mengq Y UC San Dego A53077101 mey004@eng.csd.ed Meng Xe UC San Dego A53070417 m6xe@eng.csd.ed Wenja Oyang UC San Dego A11069530 weoyang@eng.csd.ed ABSTRACT
More informationA large-alphabet oriented scheme for Chinese and English text compression
A large-alphabet orented scheme for Chnese and Englsh text compresson Hng-Yan G * Department of Compter Scence and Informaton Engneerng atonal Tawan Unersty of Scence and Technology Tape Tawan SUMMARY
More informationA Simple Methodology for Database Clustering. Hao Tang 12 Guangdong University of Technology, Guangdong, , China
for Database Clusterng Guangdong Unversty of Technology, Guangdong, 0503, Chna E-mal: 6085@qq.com Me Zhang Guangdong Unversty of Technology, Guangdong, 0503, Chna E-mal:64605455@qq.com Database clusterng
More informationLobachevsky State University of Nizhni Novgorod. Polyhedron. Quick Start Guide
Lobachevsky State Unversty of Nzhn Novgorod Polyhedron Quck Start Gude Nzhn Novgorod 2016 Contents Specfcaton of Polyhedron software... 3 Theoretcal background... 4 1. Interface of Polyhedron... 6 1.1.
More informationUB at GeoCLEF Department of Geography Abstract
UB at GeoCLEF 2006 Mguel E. Ruz (1), Stuart Shapro (2), June Abbas (1), Slva B. Southwck (1) and Davd Mark (3) State Unversty of New York at Buffalo (1) Department of Lbrary and Informaton Studes (2) Department
More informationAvailable online at Available online at Advanced in Control Engineering and Information Science
Avalable onlne at wwwscencedrectcom Avalable onlne at wwwscencedrectcom Proceda Proceda Engneerng Engneerng 00 (2011) 15000 000 (2011) 1642 1646 Proceda Engneerng wwwelsevercom/locate/proceda Advanced
More informationJournal of Chemical and Pharmaceutical Research, 2014, 6(6): Research Article. A selective ensemble classification method on microarray data
Avalable onlne www.ocpr.com Journal of Chemcal and Pharmaceutcal Research, 2014, 6(6):2860-2866 Research Artcle ISSN : 0975-7384 CODEN(USA) : JCPRC5 A selectve ensemble classfcaton method on mcroarray
More informationExperiments in Text Categorization Using Term Selection by Distance to Transition Point
Experments n Text Categorzaton Usng Term Selecton by Dstance to Transton Pont Edgar Moyotl-Hernández, Héctor Jménez-Salazar Facultad de Cencas de la Computacón, B. Unversdad Autónoma de Puebla, 14 Sur
More informationIntelligent Information Acquisition for Improved Clustering
Intellgent Informaton Acquston for Improved Clusterng Duy Vu Unversty of Texas at Austn duyvu@cs.utexas.edu Mkhal Blenko Mcrosoft Research mblenko@mcrosoft.com Prem Melvlle IBM T.J. Watson Research Center
More informationLocal Quaternary Patterns and Feature Local Quaternary Patterns
Local Quaternary Patterns and Feature Local Quaternary Patterns Jayu Gu and Chengjun Lu The Department of Computer Scence, New Jersey Insttute of Technology, Newark, NJ 0102, USA Abstract - Ths paper presents
More informationUser Authentication Based On Behavioral Mouse Dynamics Biometrics
User Authentcaton Based On Behavoral Mouse Dynamcs Bometrcs Chee-Hyung Yoon Danel Donghyun Km Department of Computer Scence Department of Computer Scence Stanford Unversty Stanford Unversty Stanford, CA
More informationHybridization of Expectation-Maximization and K-Means Algorithms for Better Clustering Performance
BULGARIAN ACADEMY OF SCIENCES CYBERNETICS AND INFORMATION TECHNOLOGIES Volume 16, No 2 Sofa 2016 Prnt ISSN: 1311-9702; Onlne ISSN: 1314-4081 DOI: 10.1515/cat-2016-0017 Hybrdzaton of Expectaton-Maxmzaton
More informationAn Optimal Algorithm to Find a Minimum 2-neighbourhood Covering Set on Cactus Graphs
Annals of Pre Appled Mathematcs Vol 2 No 1 212 45-59 ISSN: 2279-87X (P) 2279-888(onlne) Pblshed on 18 December 212 wwwresearchmathscorg Annals of An Optmal Algorthm to Fnd a Mnmm 2-neghborhood overng Set
More informationCollaboratively Regularized Nearest Points for Set Based Recognition
Academc Center for Computng and Meda Studes, Kyoto Unversty Collaboratvely Regularzed Nearest Ponts for Set Based Recognton Yang Wu, Mchhko Mnoh, Masayuk Mukunok Kyoto Unversty 9/1/013 BMVC 013 @ Brstol,
More informationClustering Algorithm of Similarity Segmentation based on Point Sorting
Internatonal onference on Logstcs Engneerng, Management and omputer Scence (LEMS 2015) lusterng Algorthm of Smlarty Segmentaton based on Pont Sortng Hanbng L, Yan Wang*, Lan Huang, Mngda L, Yng Sun, Hanyuan
More informationAlignment Results of SOBOM for OAEI 2010
Algnment Results of SOBOM for OAEI 2010 Pegang Xu, Yadong Wang, Lang Cheng, Tany Zang School of Computer Scence and Technology Harbn Insttute of Technology, Harbn, Chna pegang.xu@gmal.com, ydwang@ht.edu.cn,
More informationOutline. Discriminative classifiers for image recognition. Where in the World? A nearest neighbor recognition example 4/14/2011. CS 376 Lecture 22 1
4/14/011 Outlne Dscrmnatve classfers for mage recognton Wednesday, Aprl 13 Krsten Grauman UT-Austn Last tme: wndow-based generc obect detecton basc ppelne face detecton wth boostng as case study Today:
More informationRelevance Feedback for Image Retrieval
Vashal D Dhale et al, / (IJCSIT Internatonal Journal of Computer Scence and Informaton Technologes, Vol 4 (2, 203, 39-323 Relevance Feedback for Image Retreval Vashal D Dhale, Dr A R Mahaan, Prof Uma Thakur
More informationA Novel Adaptive Descriptor Algorithm for Ternary Pattern Textures
A Novel Adaptve Descrptor Algorthm for Ternary Pattern Textures Fahuan Hu 1,2, Guopng Lu 1 *, Zengwen Dong 1 1.School of Mechancal & Electrcal Engneerng, Nanchang Unversty, Nanchang, 330031, Chna; 2. School
More informationBioTechnology. An Indian Journal FULL PAPER. Trade Science Inc.
[Type text] [Type text] [Type text] ISSN : 0974-74 Volume 0 Issue BoTechnology 04 An Indan Journal FULL PAPER BTAIJ 0() 04 [684-689] Revew on Chna s sports ndustry fnancng market based on market -orented
More informationLoad Balancing for Hex-Cell Interconnection Network
Int. J. Communcatons, Network and System Scences,,, - Publshed Onlne Aprl n ScRes. http://www.scrp.org/journal/jcns http://dx.do.org/./jcns.. Load Balancng for Hex-Cell Interconnecton Network Saher Manaseer,
More informationFuzzy Filtering Algorithms for Image Processing: Performance Evaluation of Various Approaches
Proceedngs of the Internatonal Conference on Cognton and Recognton Fuzzy Flterng Algorthms for Image Processng: Performance Evaluaton of Varous Approaches Rajoo Pandey and Umesh Ghanekar Department of
More informationImproving KNN Method Based on Reduced Relational Grade for Microarray Missing Values Imputation
Improvng KNN Method Based on Reduced Relatonal Grade for Mcroarray Mssng Values Imputaton Yun He, De-chang P Abstract Mcroarray gene expresson data generally suffers from mssng values, whch adversely affects
More informationApproxMGMSP: A Scalable Method of Mining Approximate Multidimensional Sequential Patterns on Distributed System
ApproxMGMSP: A Scalable Method of Mnng Approxmate Multdmensonal Sequental Patterns on Dstrbuted System Changha Zhang, Kongfa Hu, Zhux Chen, Lng Chen Department of Computer Scence and Engneerng, Yangzhou
More informationOptimizing of Fuzzy C-Means Clustering Algorithm Using GA
World Acadey of Scence, Engneerng and Technology Internatonal Jornal of Copter and Inforaton Engneerng Optzng of Fzzy C-Means Clsterng Algorth Usng GA Mohanad Alata, Mohaad Molh, and Abdllah Ran Internatonal
More informationNovel Pattern-based Fingerprint Recognition Technique Using 2D Wavelet Decomposition
Mathematcal Methods for Informaton Scence and Economcs Novel Pattern-based Fngerprnt Recognton Technque Usng D Wavelet Decomposton TUDOR BARBU Insttute of Computer Scence of the Romanan Academy T. Codrescu,,
More informationPruning Training Corpus to Speedup Text Classification 1
Prunng Tranng Corpus to Speedup Text Classfcaton Jhong Guan and Shugeng Zhou School of Computer Scence, Wuhan Unversty, Wuhan, 430079, Chna hguan@wtusm.edu.cn State Key Lab of Software Engneerng, Wuhan
More informationAn Efficient Genetic Algorithm with Fuzzy c-means Clustering for Traveling Salesman Problem
An Effcent Genetc Algorthm wth Fuzzy c-means Clusterng for Travelng Salesman Problem Jong-Won Yoon and Sung-Bae Cho Dept. of Computer Scence Yonse Unversty Seoul, Korea jwyoon@sclab.yonse.ac.r, sbcho@cs.yonse.ac.r
More informationHigh-Boost Mesh Filtering for 3-D Shape Enhancement
Hgh-Boost Mesh Flterng for 3-D Shape Enhancement Hrokazu Yagou Λ Alexander Belyaev y Damng We z Λ y z ; ; Shape Modelng Laboratory, Unversty of Azu, Azu-Wakamatsu 965-8580 Japan y Computer Graphcs Group,
More informationMeta-heuristics for Multidimensional Knapsack Problems
2012 4th Internatonal Conference on Computer Research and Development IPCSIT vol.39 (2012) (2012) IACSIT Press, Sngapore Meta-heurstcs for Multdmensonal Knapsack Problems Zhbao Man + Computer Scence Department,
More informationUsing Neural Networks and Support Vector Machines in Data Mining
Usng eural etworks and Support Vector Machnes n Data Mnng RICHARD A. WASIOWSKI Computer Scence Department Calforna State Unversty Domnguez Hlls Carson, CA 90747 USA Abstract: - Multvarate data analyss
More informationThe Shortest Path of Touring Lines given in the Plane
Send Orders for Reprnts to reprnts@benthamscence.ae 262 The Open Cybernetcs & Systemcs Journal, 2015, 9, 262-267 The Shortest Path of Tourng Lnes gven n the Plane Open Access Ljuan Wang 1,2, Dandan He
More informationLearning-Based Top-N Selection Query Evaluation over Relational Databases
Learnng-Based Top-N Selecton Query Evaluaton over Relatonal Databases Lang Zhu *, Wey Meng ** * School of Mathematcs and Computer Scence, Hebe Unversty, Baodng, Hebe 071002, Chna, zhu@mal.hbu.edu.cn **
More information