CS 2710 Foundations of AI Lecture 22. Machine learning. Machine Learning

Size: px
Start display at page:

Download "CS 2710 Foundations of AI Lecture 22. Machine learning. Machine Learning"

Transcription

1 CS 7 Foudatos of AI Lecture Mache learg Mlos Hauskrecht mlos@cs.ptt.edu 539 Seott Square Mache Learg The feld of mache learg studes the desg of computer programs (agets) capable of learg from past eperece or adaptg to chages the evromet The eed for buldg agets capable of learg s everywhere predctos medce, tet ad web page classfcato, speech recogto, mage/tet retreval, commercal software

2 Learg Learg process: Learer (a computer program) processes data D represetg past epereces ad tres to ether to develop a approprate respose to future data, or descrbe some meagful way the data see Eample: Learer sees a set of patet cases (patet records) wth correspodg dagoses. It ca ether try: to predct the presece of a dsease for future patets descrbe the depedeces betwee dseases, symptoms Types of learg Supervsed learg Learg mappg betwee puts ad desred outputs y Teacher gves me y s for the learg purposes Usupervsed learg Learg relatos betwee data compoets No specfc outputs gve by a teacher Reforcemet learg Learg mappg betwee puts ad desred outputs y Crtc does ot gve me y s but stead a sgal (reforcemet) of how good my aswer was Other types of learg: Cocept learg, actve learg, deep learg,

3 Data: D { d, d,.., d d, y Supervsed learg a set of eamples s put vector, ad y s desred output (gve by a teacher) Objectve: lear the mappg f : X Y s.t. y f ( ) for all,.., Two types of problems: Regresso: X dscrete or cotuous Y s cotuous Classfcato: X dscrete or cotuous Y s dscrete } Supervsed learg eamples Regresso: Y s cotuous Debt/equty Eargs Future product orders Stock prce Data: Debt/equty Eargs Future prod orders Stock prce

4 Supervsed learg eamples Classfcato: Y s dscrete Label 3 Hadwrtte dgt (array of,s) Data: mage dgt Usupervsed learg Data: D { d, d,.., d} d vector of values No target value (output) y Objectve: lear relatos betwee samples, compoets of samples Types of problems: Clusterg Group together smlar eamples, e.g. patet cases Desty estmato Model probablstcally the populato of samples

5 Usupervsed learg eample Clusterg. Group together smlar eamples d Usupervsed learg eample Clusterg. Group together smlar eamples d

6 Usupervsed learg eample Desty estmato. We wat to buld a probablty model P() of a populato from whch we drew eamples d Usupervsed learg. Desty estmato A probablty desty of a pot the two dmesoal space Model used here: Mture of Gaussas

7 Reforcemet learg We wat to lear: f : X Y We see eamples of puts but ot y We select y for observed from avalable choces We get a feedback (reforcemet) from a crtc about how good our choce of y was Iput Learer Output/acto y reforcemet Crtc The goal s to select outputs that lead to the best reforcemet Learg: frst look Assume we see eamples of pars (, y) D ad we wat to lear the mappg f : X Y to predct y for some future We get the data D - what should we do? y

8 Learg: frst look Problem: may possble fuctos f : X Y ests for represetg the mappg betwee ad y Whch oe to choose? May eamples stll usee! y Learg: frst look Soluto: make a assumpto about the model, say, f ( ) a b y

9 Learg: frst look Choosg a parametrc model or a set of models s ot eough Stll too may fuctos f ( ) a b Oe for every par of parameters a, b y Learg: frst look We wat the best set of model parameters reduce the msft betwee the model M ad observed data D Or, ( other words) epla the data the best How to measure the msft? y

10 Learg: frst look We wat the best set of model parameters reduce the msft betwee the model M ad observed data D Or, ( other words) epla the data the best How to measure the msft? y - The dfferece observed value of y ad model predcto Learg: frst look We wat the best set of model parameters reduce the msft betwee the model M ad observed data D Or, ( other words) epla the data the best How to measure the msft? y

11 Learg: frst look We wat the best set of model parameters reduce the msft betwee the model M ad observed data D Or, ( other words) epla the data the best How to measure the msft? Objectve fucto: Error fucto: Measures the msft betwee D ad M Eamples of error fuctos: Average Square Error Average Absolute Error ( y y f ( )) f ( ) Learg: frst look Lear regresso problem Mmzes the squared error fucto for the lear model ( y f ( )) y

12 Learg: frst look Applcato: A ew eample wth ukow value y s checked agast the model, ad y s calculated as y f ( ) a b yy y =? Supervsed learg: Classfcato Data D: pars (, y) where y s a class label: y eamples: patet wll be readmtted or o, has dsease (case) or o (cotrol) case cotrol

13 Supervsed learg: Classfcato Fd a model f: X R, say f ( ) a b c that defes a decso boudary f () = that separates well the two classes Note that some eamples are ot correctly classfed case f () = cotrol Supervsed learg: Classfcato A ew eample wth ukow class label s checked agast the model, the class label s assged case cotrol? 3

14 Learg: frst look. Data: D { d, d,.., d}. Model selecto: Select a model or a set of models (wth parameters) E.g. y a b 3. Choose the objectve fucto Squared error ( y f ( )). Learg: Fd the set of parameters optmzg the error fucto The model ad parameters wth the smallest error 5. Applcato Apply the leared model to ew data E.g. predct ys for ew puts usg leared f () CS 75 Mache Learg Learg: frst look. Data: D { d, d,.., d}. Model selecto: Select a model or a set of models (wth parameters) E.g. y a y b 3. Choose the objectve fucto Squared error ( y f ( )) -. Learg: Fd the set of parameters optmzg the error fucto - The model ad parameters wth the smallest error 5. Applcato: Apply the leared model to ew data E.g. predct ys for ew puts usg leared f () CS 75 Mache Learg

15 A learg system: basc cycle. Data: D { d, d,.., d}. Model selecto: Select a model or a set of models (wth parameters) E.g. y a b 3. Choose the objectve fucto Squared error y ( y f ( )). Learg: Fd the set of parameters optmzg the error fucto The model ad parameters wth the smallest error - 5. Applcato Apply the leared model to ew data - E.g. predct ys for ew puts usg leared f () CS 75 Mache Learg Learg: frst look. Data: D { d, d,.., d}. Model selecto: Select a model or a set of models (wth parameters) E.g. y a b 3. Choose the objectve fucto Squared error ( y f ( )). Learg: Fd the set of parameters optmzg the error fucto The model ad parameters wth the smallest error 5. Applcato Apply the leared - model to ew data E.g. predct ys for - ew puts usg leared f () CS 75 Mache Learg 5

16 Learg: frst look. Data: D { d, d,.., d}. Model selecto: Select a modelor a set of models (wth parameters) - E.g. y a b - 3. Choose the objectve fucto Squared error ( y f ( )). Learg: Fd the set of parameters optmzg the error fucto The model ad parameters wth the smallest error 5. Applcato: Apply the leared model to ew data E.g. predct ys for ew puts usg leared f () Learg: frst look. Data: D { d, d,.., d}. Model selecto: Select a model or a set of models (wth parameters) E.g. y a b 3. Choose the objectve fucto Squared error ( y f ( )). Learg: - Fd the set of parameters - optmzg the error fucto The model ad - parameters wth the smallest error Applcato: Apply the leared model to ew data E.g. predct ys for ew puts usg leared f ()

17 Learg: frst look. Data: D { d, d,.., d}. Model selecto: Select a model or a set of models (wth parameters) E.g. y a b 3. Choose the objectve fucto Squared error ( y f ( )). Learg: Fd the set of parameters optmzg the error fucto The model ad parameters wth the smallest error 5. Applcato Apply the leared model to ew data Looks straghtforward, but there are problems. Learg Problem We ft the model based o past eperece (past eamples see) But ultmately we are terested learg the mappg that performs well o the whole populato of eamples Trag data: Data used to ft the parameters of the model Trag error: ( y f ( )) True (geeralzato) error (over the whole ukow populato): E(, y )( y f ( )) Epected squared error Trag error tres to appromate the true error!!!! Does a good trag error mply a good geeralzato error? 7

18 Overfttg Assume we have a set of pots ad we cosder polyomal fuctos as our possble models Overfttg Fttg a lear fucto wth the square error Error s ozero

19 Overfttg Fttg a lear fucto wth mea-squares error Error s ozero Overfttg Lear vs. cubc polyomal Hgher order polyomal leads to a better ft, smaller error

20 Overfttg Is t always good to mmze the error of the observed data? Overfttg For data pots, degree 9 polyomal gves a perfect ft (Lagrage terpolato). Error s zero. Is t always good to mmze the trag error?

21 Overfttg For data pots, degree 9 polyomal gves a perfect ft (Lagrage terpolato). Error s zero. Is t always good to mmze the trag error? NO!! More mportat: How do we perform o the usee data? Overfttg Stuato whe the trag error s low ad the geeralzato error s hgh. Causes of the pheomeo: Model wth more degrees of freedom (more parameters) Small data sze (as compared to the complety of model)

22 How to evaluate the learer s performace? Geeralzato error s the true error for the populato of eamples we would lke to optmze E (, y )( y f ( )) But t caot be computed eactly Optmzg (mea) trag error ca lead to overft,.e. trag error may ot reflect properly the geeralzato error ( y f ( )),.. So how to test the geeralzato error? How to assess the learer s performace? Geeralzato error s the true error for the populato of eamples we would lke to optmze E [( y f ( )) ] (, y ) Sample mea oly appromates t How to measure the geeralzato error? Two ways: Theoretcal: Law of Large umbers statstcal bouds o the dfferece betwee the true ad sample mea errors Practcal: Use a separate data set wth m data samples to test (Mea) test error ( y j f ( j )) m j,.. m

23 Testg of learg models Smple holdout method Dvde the data to the trag ad test data Dataset Trag set Testg set Evaluate Lear (ft) Predctve model Typcally /3 trag ad /3 testg Testg of models Data set Trag set Test set case cotrol case cotrol Lear o the trag set The model Evaluate o the test set 3

24 Data: Desty estmato D { D, D,.., D} D a vector of attrbute values Objectve: estmate the model of the uderlyg probablty dstrbuto over varables X, p(x), usg eamples D true dstrbuto samples p (X) D D, D,.., D } { estmate pˆ ( X) Desty estmato true dstrbuto samples p (X) D D, D,.., D } { estmate pˆ ( X) Stadard (d) assumptos: Samples are depedet of each other come from the same (detcal) dstrbuto (fed p(x) ) Idepedetly draw staces from the same fed dstrbuto

25 Learg va parameter estmato I ths lecture we cosder parametrc desty estmato Basc settgs: A set of radom varables X { X, X,, Xd} A model of the dstrbuto over varables X wth parameters Data D D, D,.., D } { Objectve: fd parameters ˆ that ft the data the best What s the best set of parameters? There are varous crtera oe ca apply here. Parameter estmato. Basc crtera. Mamum lkelhood (ML) mamze p( D, ) - represets pror (backgroud) kowledge Mamum a posteror probablty (MAP) mamze p( D, ) Selects the mode of the posteror p( D, ) p( D, ) p( ) p( D ) 5

Machine Learning. CS 2750 Machine Learning. Administration. Lecture 1. Milos Hauskrecht 5329 Sennott Square, x4-8845

Machine Learning. CS 2750 Machine Learning. Administration. Lecture 1. Milos Hauskrecht 5329 Sennott Square, x4-8845 CS 75 Mache Learg Lecture Mache Learg Mlos Hauskrecht mlos@cs.ptt.edu 539 Seott Square, 5 people.cs.ptt.edu/~mlos/courses/cs75/ Admstrato Istructor: Prof. Mlos Hauskrecht mlos@cs.ptt.edu 539 Seott Square,

More information

Machine Learning: Algorithms and Applications

Machine Learning: Algorithms and Applications /03/ Mache Learg: Algorthms ad Applcatos Florao Z Free Uversty of Boze-Bolzao Faculty of Computer Scece Academc Year 0-0 Lecture 3: th March 0 Naïve Bayes classfer ( Problem defto A trag set X, where each

More information

Multiclass classification Decision trees

Multiclass classification Decision trees CS 75 Mache Learg Lecture Multclass classfcato Decso trees Mlos Hauskrecht mlos@cs.tt.edu 59 Seott Suare CS 75 Mache Learg Mdterm eam Mdterm Tuesda, March 4, 4 I-class 75 mutes closed book materal covered

More information

Point Estimation-III: General Methods for Obtaining Estimators

Point Estimation-III: General Methods for Obtaining Estimators Pot Estmato-III: Geeral Methods for Obtag Estmators RECAP 0.-0.6 Data: Radom Sample from a Populato of terest o Real valued measuremets: o Assumpto (Hopefully Reasoable) o Model: Specfed Probablty Dstrbuto

More information

Designing a learning system

Designing a learning system CS 75 Mache Leafrg Lecture 3 Desgg a learg system Mlos Hauskrecht mlos@cs.ptt.edu 539 Seott Square, x4-8845 people.cs.ptt.edu/~mlos/courses/cs75/ Homeork assgmet Homeork assgmet ll be out today To parts:

More information

Designing a learning system

Designing a learning system CS 75 Itro to Machie Learig Lecture Desigig a learig system Milos Hauskrecht milos@pitt.edu 539 Seott Square, -5 people.cs.pitt.edu/~milos/courses/cs75/ Admiistrivia No homework assigmet this week Please

More information

Fitting. We ve learned how to detect edges, corners, blobs. Now what? We would like to form a. compact representation of

Fitting. We ve learned how to detect edges, corners, blobs. Now what? We would like to form a. compact representation of Fttg Fttg We ve leared how to detect edges, corers, blobs. Now what? We would lke to form a hgher-level, h l more compact represetato of the features the mage b groupg multple features accordg to a smple

More information

Bezier curves. 1. Defining a Bezier curve. A closed Bezier curve can simply be generated by closing its characteristic polygon

Bezier curves. 1. Defining a Bezier curve. A closed Bezier curve can simply be generated by closing its characteristic polygon Curve represetato Copyrght@, YZU Optmal Desg Laboratory. All rghts reserved. Last updated: Yeh-Lag Hsu (--). Note: Ths s the course materal for ME55 Geometrc modelg ad computer graphcs, Yua Ze Uversty.

More information

Face Recognition using Supervised & Unsupervised Techniques

Face Recognition using Supervised & Unsupervised Techniques Natoal Uversty of Sgapore EE5907-Patter recogto-2 NAIONAL UNIVERSIY OF SINGAPORE EE5907 Patter Recogto Project Part-2 Face Recogto usg Supervsed & Usupervsed echques SUBMIED BY: SUDEN NAME: harapa Reddy

More information

Designing a learning system

Designing a learning system CS 75 Machie Learig Lecture Desigig a learig system Milos Hauskrecht milos@cs.pitt.edu 539 Seott Square, x-5 people.cs.pitt.edu/~milos/courses/cs75/ Admiistrivia No homework assigmet this week Please try

More information

ANALYSIS OF VARIANCE WITH PARETO DATA

ANALYSIS OF VARIANCE WITH PARETO DATA Proceedgs of the th Aual Coferece of Asa Pacfc Decso Sceces Isttute Hog Kog, Jue -8, 006, pp. 599-609. ANALYSIS OF VARIANCE WITH PARETO DATA Lakhaa Watthaacheewakul Departmet of Mathematcs ad Statstcs,

More information

Software reliability is defined as the probability of failure

Software reliability is defined as the probability of failure Evolutoary Regresso Predcto for Software Cumulatve Falure Modelg: a comparatve study M. Beaddy, M. Wakrm & S. Aljahdal 2 : Dept. of Math. & Ifo. Equpe MMS, Ib Zohr Uversty Morocco. beaddym@yahoo.fr 2:

More information

ChEn 475 Statistical Analysis of Regression Lesson 1. The Need for Statistical Analysis of Regression

ChEn 475 Statistical Analysis of Regression Lesson 1. The Need for Statistical Analysis of Regression Statstcal-Regresso_hadout.xmcd Statstcal Aalss of Regresso ChE 475 Statstcal Aalss of Regresso Lesso. The Need for Statstcal Aalss of Regresso What do ou do wth dvdual expermetal data pots? How are the

More information

Unsupervised Discretization Using Kernel Density Estimation

Unsupervised Discretization Using Kernel Density Estimation Usupervsed Dscretzato Usg Kerel Desty Estmato Maregle Bba, Floraa Esposto, Stefao Ferll, Ncola D Mauro, Teresa M.A Basle Departmet of Computer Scece, Uversty of Bar Va Oraboa 4, 7025 Bar, Italy {bba,esposto,ferll,dm,basle}@d.uba.t

More information

Active Bayesian Learning For Mixture Models

Active Bayesian Learning For Mixture Models Actve Bayesa Learg For Mxture Models Ia Davdso Slco Graphcs 300 Crttede L, MS 876 Mouta Vew, CA 94587 pd@hotmal.com Abstract Tradtoally, Bayesa ductve learg volves fdg the most probable model from the

More information

Descriptive Statistics: Measures of Center

Descriptive Statistics: Measures of Center Secto 2.3 Descrptve Statstcs: Measures of Ceter Frequec dstrbutos are helpful provdg formato about categorcal data, but wth umercal data we ma wat more formato. Statstc: s a umercal measure calculated

More information

Chapter 3 Descriptive Statistics Numerical Summaries

Chapter 3 Descriptive Statistics Numerical Summaries Secto 3.1 Chapter 3 Descrptve Statstcs umercal Summares Measures of Cetral Tedecy 1. Mea (Also called the Arthmetc Mea) The mea of a data set s the sum of the observatos dvded by the umber of observatos.

More information

Prof. Feng Liu. Winter /24/2019

Prof. Feng Liu. Winter /24/2019 Prof. Feg Lu Wter 209 http://www.cs.pd.edu/~flu/courses/cs40/ 0/24/209 Last Tme Feature detecto 2 Toda Feature matchg Fttg The followg sldes are largel from Prof. S. Lazebk 3 Wh etract features? Motvato:

More information

2 General Regression Neural Network (GRNN)

2 General Regression Neural Network (GRNN) 4 Geeral Regresso Neural Network (GRNN) GRNN, as proposed b oald F. Specht [Specht 9] falls to the categor of probablstc eural etworks as dscussed Chapter oe. Ths eural etwork lke other probablstc eural

More information

For all questions, answer choice E) NOTA" means none of the above answers is correct. A) 50,500 B) 500,000 C) 500,500 D) 1,001,000 E) NOTA

For all questions, answer choice E) NOTA means none of the above answers is correct. A) 50,500 B) 500,000 C) 500,500 D) 1,001,000 E) NOTA For all questos, aswer choce " meas oe of the above aswers s correct.. What s the sum of the frst 000 postve tegers? A) 50,500 B) 500,000 C) 500,500 D),00,000. What s the sum of the tegers betwee 00 ad

More information

APPLICATION OF CLUSTERING METHODS IN BANK S PROPENSITY MODEL

APPLICATION OF CLUSTERING METHODS IN BANK S PROPENSITY MODEL APPLICATION OF CLUSTERING METHODS IN BANK S PROPENSITY MODEL Sergej Srota Haa Řezaková Abstract Bak s propesty models are beg developed for busess support. They should help to choose clets wth a hgher

More information

ABSTRACT Keywords

ABSTRACT Keywords A Preprocessg Scheme for Hgh-Cardalty Categorcal Attrbutes Classfcato ad Predcto Problems Daele Mcc-Barreca ClearCommerce Corporato 1100 Metrc Blvd. Aust, TX 78732 ABSTRACT Categorcal data felds characterzed

More information

A genetic procedure used to train RFB neural networks

A genetic procedure used to train RFB neural networks A geetc procedure used to tra RFB eural etworks Costat-Iula VIZITIU Commucatos ad Electroc Systems Departmet Mltary Techcal Academy George Cosbuc Aveue 8-83 5 th Dstrct Bucharest ROMANIA vc@mta.ro http://www.mta.ro

More information

Signal Classification Method Based on Support Vector Machine and High-Order Cumulants

Signal Classification Method Based on Support Vector Machine and High-Order Cumulants Wreless Sesor Network,,, 48-5 do:.46/ws..7 Publshed Ole Jauary (http://www.scrp.org/joural/ws/). Sgal Classfcato Method Based o Support Vector Mache ad Hgh-Order Cumulats Abstract X ZHOU, Yg WU, B YANG

More information

LP: example of formulations

LP: example of formulations LP: eample of formulatos Three classcal decso problems OR: Trasportato problem Product-m problem Producto plag problem Operatos Research Massmo Paolucc DIBRIS Uversty of Geova Trasportato problem The decso

More information

Blind Steganalysis for Digital Images using Support Vector Machine Method

Blind Steganalysis for Digital Images using Support Vector Machine Method 06 Iteratoal Symposum o Electrocs ad Smart Devces (ISESD) November 9-30, 06 Bld Stegaalyss for Dgtal Images usg Support Vector Mache Method Marcelus Hery Meor School of Electrcal Egeerg ad Iformatcs Badug

More information

Using Linear-threshold Algorithms to Combine Multi-class Sub-experts

Using Linear-threshold Algorithms to Combine Multi-class Sub-experts Usg Lear-threshold Algorthms to Combe Mult-class Sub-experts Chrs Mesterharm MESTERHA@CS.RUTGERS.EDU Rutgers Computer Scece Departmet 110 Frelghuyse Road Pscataway, NJ 08854 USA Abstract We preset a ew

More information

Spatial Interpolation Using Neural Fuzzy Technique

Spatial Interpolation Using Neural Fuzzy Technique Wog, K.W., Gedeo, T., Fug, C.C. ad Wog, P.M. (00) Spatal terpolato usg eural fuzzy techque. I: Proceedgs of the 8th Iteratoal Coferece o Neural Iformato Processg (ICONIP), Shagha, Cha Spatal Iterpolato

More information

A Comparison of Univariate Smoothing Models: Application to Heart Rate Data Marcus Beal, Member, IEEE

A Comparison of Univariate Smoothing Models: Application to Heart Rate Data Marcus Beal, Member, IEEE A Comparso of Uvarate Smoothg Models: Applcato to Heart Rate Data Marcus Beal, Member, IEEE E-mal: bealm@pdx.edu Abstract There are a umber of uvarate smoothg models that ca be appled to a varety of olear

More information

MATHEMATICAL PROGRAMMING MODEL OF THE CRITICAL CHAIN METHOD

MATHEMATICAL PROGRAMMING MODEL OF THE CRITICAL CHAIN METHOD MATHEMATICAL PROGRAMMING MODEL OF THE CRITICAL CHAIN METHOD TOMÁŠ ŠUBRT, PAVLÍNA LANGROVÁ CUA, SLOVAKIA Abstract Curretly there s creasgly dcated that most of classcal project maagemet methods s ot sutable

More information

Electrocardiogram Classification Method Based on SVM

Electrocardiogram Classification Method Based on SVM Electrocardogram Classfcato Method Based o SVM Xao Tag Zhwe Mo College of mathematcs ad software scece, Schua ormal uversty, Chegdu 60066, P. R. Cha Abstract Heart dsease s oe of the ma dseases threateg

More information

Learning Graphical Models from a Distributed Stream

Learning Graphical Models from a Distributed Stream Learg Graphcal Models from a Dstrbuted Stream Yu Zhag #1, Srkata Trthapura #2, Graham Cormode # Electrcal ad Computer Egeerg Departmet, Iowa State Uversty 1 yuz1988@astate.edu 2 st@astate.edu Uversty of

More information

Biological Neurons. Biological Neuron: Information Processing. Axon Hillock. Axon. Soma. Nucleus. Dendrite. Terminal. M. A. El Sharkawi, NN General 2

Biological Neurons. Biological Neuron: Information Processing. Axon Hillock. Axon. Soma. Nucleus. Dendrite. Terminal. M. A. El Sharkawi, NN General 2 Bolocal Neuros Soa Ao Hlloc Ao Nucleus Dedrte eral M. A. El Shara, NN Geeral Bolocal Neuro: Iforato Process M. A. El Shara, NN Geeral 3 M. A. El Shara, NN Geeral 4 Recoto of Bolocal Neuros Lear (tra Recall

More information

Keywords Classification, Texture Characterization, LBP, FLBP, Medical Imaging

Keywords Classification, Texture Characterization, LBP, FLBP, Medical Imaging Volume 4, Issue 8, August 04 ISSN: 77 8X Iteratoal Joural of Advaced Research Computer Scece ad Software Egeerg Research Paper Avalable ole at: www.jarcsse.com Dagose the Thyrod Usg Teture Characterzato

More information

DESIGN AND EVALUATION OF EXPERIMENTS WITH SAS

DESIGN AND EVALUATION OF EXPERIMENTS WITH SAS XIX IMEKO World Cogress Fudametal ad ppled Metrology September 6, 009, Lsbo, Portugal DESIGN ND EVLUTION OF EXPERIMENTS WITH draa Horíová Uversty of Ecoomcs Bratslava (Faculty of Ecoomcs Iformatcs, Isttute

More information

Transistor/Gate Sizing Optimization

Transistor/Gate Sizing Optimization Trasstor/Gate Szg Optmzato Gve: Logc etwork wth or wthout cell lbrary Fd: Optmal sze for each trasstor/gate to mmze area or power, both uder delay costrat Statc szg: based o tmg aalyss ad cosder all paths

More information

Reliable Surface Extraction from Point-Clouds using Scanner-Dependent Parameters

Reliable Surface Extraction from Point-Clouds using Scanner-Dependent Parameters 1 Relable Surface Extracto from Pot-Clouds usg Scaer-Depedet Parameters Hrosh Masuda 1, Ichro Taaka 2, ad Masakazu Eomoto 3 1 The Uversty of Tokyo, masuda@sys.t.u-tokyo.ac.jp 2 Tokyo Dek Uversty, taaka@cck.deda.ac.jp

More information

Clustering documents with vector space model using n-grams

Clustering documents with vector space model using n-grams Clusterg documets wth vector space model usg -grams Klas Skogmar, d97ksk@efd.lth.se Joha Olsso, d97jo@efd.lth.se Lud Isttute of Techology Supervsed by: Perre Nugues, Perre.Nugues@cs.lth.se Abstract Ths

More information

Integration of Support Vector Machine and Bayesian Neural Network for Data Mining and Classification

Integration of Support Vector Machine and Bayesian Neural Network for Data Mining and Classification World Academy of Scece, Egeerg ad Techology Iteratoal Joural of Computer ad Iformato Egeerg Vol:4, No:4, 2 Itegrato of Support Vector Mache ad Bayesa Neural Network for Data Mg ad Classfcato Essam Al-Daoud

More information

SVM Classification Method Based Marginal Points of Representative Sample Sets

SVM Classification Method Based Marginal Points of Representative Sample Sets P P College P P College P Iteratoal Joural of Iformato Techology Vol. No. 9 005 SVM Classfcato Method Based Margal Pots of Represetatve Sample Sets Wecag ZhaoP P, Guagrog JP P, Ru NaP P, ad Che FegP of

More information

Pattern Extraction, Classification and Comparison Between Attribute Selection Measures

Pattern Extraction, Classification and Comparison Between Attribute Selection Measures Subrata Pramak et al. / (IJCSIT) Iteratoal Joural of Computer Scece ad Iformato Techoes, Vol. 1 (5), 010, 371-375 Patter Extracto, Classfcato ad Comparso Betwee ttrbute Selecto Measures Subrata Pramak,

More information

FEATURE SELECTION ON COMBINATIONS FOR EFFICIENT LEARNING FROM IMAGES. Rong Xiao, Lei Zhang, and Hong-Jiang Zhang

FEATURE SELECTION ON COMBINATIONS FOR EFFICIENT LEARNING FROM IMAGES. Rong Xiao, Lei Zhang, and Hong-Jiang Zhang FEATURE SELECTION ON COMBINATIONS FOR EFFICIENT LEARNING FROM IMAGES Rog Xao, Le Zhag, ad Hog-Jag Zhag Mcrosoft Research Asa, Bejg 100080, P.R. Cha {t-rxao, lezhag, hjzhag}@mcrosoft.com ABSTRACT Due to

More information

Estimating Feasibility Using Multiple Surrogates and ROC Curves

Estimating Feasibility Using Multiple Surrogates and ROC Curves Estmatg Feasblty Usg Multple Surrogates ad ROC Curves Arba Chaudhur * Uversty of Florda, Gaesvlle, Florda, 3601 Rodolphe Le Rche École Natoale Supéreure des Mes de Sat-Étee, Sat-Étee, Frace ad CNRS LIMOS

More information

Multiscale Principal Component Analysis

Multiscale Principal Component Analysis Multscale Prcpal Compoet Aalyss Thess submtted for the degree of Doctor of Phlosophy at the uversty of Lecester By Akduko Ayode Akwum Departmet of Mathematcs Uversty of Lecester August 05 Abstract The

More information

ECE Digital Image Processing and Introduction to Computer Vision

ECE Digital Image Processing and Introduction to Computer Vision ECE59064 Dgtal Image Processg ad Itroducto to Computer Vso Depart. of ECE NC State Uverst Istructor: Tafu Matt Wu Sprg 07 Outle Recap Le Segmet Detecto Fttg Least square Total square Robust estmator Hough

More information

A Perception of Statistical Inference in Data Mining

A Perception of Statistical Inference in Data Mining Iteratoal Joural of Computer Scece & Commucato Vol., No., July-December 00, pp. 373-378 A Percepto of Statstcal Iferece Data Mg Sajay Gaur & M. S. Dulawat, Departmet of Mathematcs & Statstcs, Mohalal Sukhada

More information

APR 1965 Aggregation Methodology

APR 1965 Aggregation Methodology Sa Joaqu Valley Ar Polluto Cotrol Dstrct APR 1965 Aggregato Methodology Approved By: Sged Date: March 3, 2016 Araud Marjollet, Drector of Permt Servces Backgroud Health rsk modelg ad the collecto of emssos

More information

Office Hours. COS 341 Discrete Math. Office Hours. Homework 8. Currently, my office hours are on Friday, from 2:30 to 3:30.

Office Hours. COS 341 Discrete Math. Office Hours. Homework 8. Currently, my office hours are on Friday, from 2:30 to 3:30. Oce Hours Curretly, my oce hours are o Frday, rom :30 to 3:30. COS 31 Dscrete Math 1 Oce Hours Curretly, my oce hours are o Frday, rom :30 to 3:30. Nobody seems to care. Chage oce hours? Tuesday, 8 PM

More information

Fingerprint Classification Based on Spectral Features

Fingerprint Classification Based on Spectral Features Fgerprt Classfcato Based o Spectral Features Hosse Pourghassem Tarbat Modares Uversty h_poorghasem@modares.ac.r Hassa Ghassema Tarbat Modares Uversty ghassem@modares.ac.r Abstract: Fgerprt s oe of the

More information

Estimation of Co-efficient of Variation in PPS sampling.

Estimation of Co-efficient of Variation in PPS sampling. It. Statstcal Ist.: Proc. 58th World Statstcal Cogress, 0, Dubl (Sesso CPS00) p.409 Estmato of Co-effcet of Varato PPS samplg. Archaa. V ( st Author) Departmet of Statstcs, Magalore Uverst Magalagagotr,

More information

Application Research for Ultrasonic Flaw Identification Based on Support Vector Machine Jing Huang 1, a, Binglei Guan 1, b

Application Research for Ultrasonic Flaw Identification Based on Support Vector Machine Jing Huang 1, a, Binglei Guan 1, b 4th Iteratoal Coferece o Mechatrocs, Materals, Chemstry ad Computer Egeerg (ICMMCCE 205) Applcato Research for Ultrasoc Flaw Idetfcato Based o Support Vector Mache Jg Huag, a, Bgle Gua, b School of Electroc

More information

Unsupervised visual learning of three-dimensional objects using a modular network architecture

Unsupervised visual learning of three-dimensional objects using a modular network architecture PERGAMON Neural Networks 12 (1999) 1037 1051 Neural Networks www.elsever.com/locate/euet Usupervsed vsual learg of three-dmesoal objects usg a modular etwork archtecture H. Ado a, *, S. Suzuk a,b, T. Fujta

More information

1-D matrix method. U 4 transmitted. incident U 2. reflected U 1 U 5 U 3 L 2 L 3 L 4. EE 439 matrix method 1

1-D matrix method. U 4 transmitted. incident U 2. reflected U 1 U 5 U 3 L 2 L 3 L 4. EE 439 matrix method 1 -D matrx method We ca expad the smple plae-wave scatterg for -D examples that we ve see to a more versatle matrx approach that ca be used to hadle may terestg -D problems. The basc dea s that we ca break

More information

Prediction Method of Network Security Situation Based on GA- LSSVM Time Series Analysis

Prediction Method of Network Security Situation Based on GA- LSSVM Time Series Analysis AMSE JOURNALS-AMSE IIETA publcato-07-seres: Advaces B; Vol. 60; N ; pp 37-390 Submtted Mar. 05 07; Revsed May 5 07; Accepted Ju. 0 07 Predcto Method of Network Securty Stuato Based o GA- LSSVM Tme Seres

More information

Network Intrusion Detection Using Support Vector Machine Based on Particle Swarm Optimization Li Wang 1,2, Chunhua Dong 2, Jianping Hu 2, Guodong Li 2

Network Intrusion Detection Using Support Vector Machine Based on Particle Swarm Optimization Li Wang 1,2, Chunhua Dong 2, Jianping Hu 2, Guodong Li 2 Iteratoal Coferece o Appled Scece ad Egeerg Iovato (ASEI 015) Network Itruso Detecto Usg Support Vector Mache Based o Partcle Swarm Optmzato L Wag 1,, Chuhua Dog, Japg Hu, Guodog L 1.School of Electrocs

More information

Comparison Studies on Classification for Remote Sensing Image Based on Data Mining Method

Comparison Studies on Classification for Remote Sensing Image Based on Data Mining Method Hag Xao ad Xub Zhag Comparso Studes o Classfcato for Remote Sesg Image Based o Data Mg Method Hag XIAO 1, Xub ZHANG 1 1: School of Electroc, Iformato ad Electrcal Egeerg Shagha Jaotog Uversty No. 1954,

More information

A SVM Gray-Box Model for a Solid Substrate Fermentation Process

A SVM Gray-Box Model for a Solid Substrate Fermentation Process A publcato of CHEMICAL ENGINEERING TRANSACTIONS VOL. 35, 2013 Guest Edtors: Petar Varbaov, Jří Klemeš, Paos Seferls, Athaasos I. Papadopoulos, Spyros Voutetaks Copyrght 2013, AIDIC Servz S.r.l., ISBN 978-88-95608-26-6;

More information

Optimal Allocation of Complex Equipment System Maintainability

Optimal Allocation of Complex Equipment System Maintainability Optmal Allocato of Complex Equpmet System ataablty X Re Graduate School, Natoal Defese Uversty, Bejg, 100091, Cha edcal Protecto Laboratory, Naval edcal Research Isttute, Shagha, 200433, Cha Emal:rexs841013@163.com

More information

Construction Project Cost Prediction Based on Genetic Algorithm and Least Squares Support Vector Machine

Construction Project Cost Prediction Based on Genetic Algorithm and Least Squares Support Vector Machine 5th Iteratoal Coferece o Cvl Egeerg ad rasportato (ICCE 05) Costructo Project Cost Predcto Based o Geetc Algorthm ad Least Squares Support Vector Mache Mg Xu, a, Bgfeg Xu,b*, Lajag Zhou,c ad L Wu,d School

More information

Regression Analysis. Acknowledgments

Regression Analysis. Acknowledgments PT 3 - Lear Regresso Regresso Aalyss How to develop ad assess a CER All models are wrog, but some are useful. -George Box I mathematcs, cotext obscures structure. I data aalyss, cotext provdes meag. -George

More information

Marcus Gallagher School of Information Technology and Electrical Engineering The University of Queensland QLD 4072, Australia

Marcus Gallagher School of Information Technology and Electrical Engineering The University of Queensland QLD 4072, Australia O the Importace of Dversty Mateace Estmato of Dstrbuto Algorthms Bo Yua School of Iformato Techology ad Electrcal Egeerg The Uversty of Queeslad QLD 4072, Australa +6-7-3365636 boyua@tee.uq.edu.au Marcus

More information

Evolutionary Strategies for Multi-Scale Radial Basis Function Kernels in Support Vector Machines

Evolutionary Strategies for Multi-Scale Radial Basis Function Kernels in Support Vector Machines Eolutoary Strateges for Mult-Scale Radal Bass Fucto Kerels Support Vector Maches Taasaee Phethrakul Departmet of Computer Egeerg Faculty of Egeerg, Chulalogkor Uersty Bagkok, Thalad 0330 taasaee@yahoo.com

More information

COMSC 2613 Summer 2000

COMSC 2613 Summer 2000 Programmg II Fal Exam COMSC 63 Summer Istructos: Name:. Prt your ame the space provded Studet Id:. Prt your studet detfer the space Secto: provded. Date: 3. Prt the secto umber of the secto whch you are

More information

SALAM A. ISMAEEL Computer Man College for Computer Studies, Khartoum / Sudan

SALAM A. ISMAEEL Computer Man College for Computer Studies, Khartoum / Sudan AAPTIVE HYBRI-WAVELET ETHO FOR GPS/ SYSTE INTEGRATION SALA A. ISAEEL Computer a College for Computer Studes, Khartoum / Suda salam.smaeel@gmal.com ABSTRACT I ths paper, a techque for estmato a global postog

More information

Journal of Chemical and Pharmaceutical Research, 2015, 7(3): Research Article

Journal of Chemical and Pharmaceutical Research, 2015, 7(3): Research Article Avalable ole www.ocpr.com Joural of Chemcal ad Pharmaceutcal Research, 2015, 73):476-481 Research Artcle ISSN : 0975-7384 CODENUSA) : JCPRC5 Research o cocept smlarty calculato method based o sematc grd

More information

Constructive Semi-Supervised Classification Algorithm and Its Implement in Data Mining

Constructive Semi-Supervised Classification Algorithm and Its Implement in Data Mining Costructve Sem-Supervsed Classfcato Algorthm ad Its Implemet Data Mg Arvd Sgh Chadel, Arua Twar, ad Naredra S. Chaudhar Departmet of Computer Egg. Shr GS Ist of Tech.& Sc. SGSITS, 3, Par Road, Idore (M.P.)

More information

International Mathematical Forum, 1, 2006, no. 31, ON JONES POLYNOMIALS OF GRAPHS OF TORUS KNOTS K (2, q ) Tamer UGUR, Abdullah KOPUZLU

International Mathematical Forum, 1, 2006, no. 31, ON JONES POLYNOMIALS OF GRAPHS OF TORUS KNOTS K (2, q ) Tamer UGUR, Abdullah KOPUZLU Iteratoal Mathematcal Forum,, 6, o., 57-54 ON JONES POLYNOMIALS OF RAPHS OF TORUS KNOTS K (, q ) Tamer UUR, Abdullah KOPUZLU Atatürk Uverst Scece Facult Dept. of. Math. 54 Erzurum, Turkey tugur@atau.edu.tr

More information

NEURO FUZZY MODELING OF CONTROL SYSTEMS

NEURO FUZZY MODELING OF CONTROL SYSTEMS NEURO FUZZY MODELING OF CONTROL SYSTEMS Efré Gorrosteta, Carlos Pedraza Cetro de Igeería y Desarrollo Idustral CIDESI, Av Pe de La Cuesta 70. Des. Sa Pablo. Querétaro, Qro, Méxco gorrosteta@teso.mx pedraza@cdes.mx

More information

Approximation of Curves Contained on the Surface by Freed-Forward Neural Networks

Approximation of Curves Contained on the Surface by Freed-Forward Neural Networks Appromato of Curves Cotaed o the Surface by Freed-Forward Neural Networks Zheghua Zhou ad Jawe Zhao Departmet of formato ad mathematcs Sceces Cha Jlag Uversty, Hagzhou, 3008, Cha zzhzjw003@63.com Abstract.

More information

Using The ACO Algorithm in Image Segmentation for Optimal Thresholding 陳香伶財務金融系

Using The ACO Algorithm in Image Segmentation for Optimal Thresholding 陳香伶財務金融系 教專研 95P- Usg The ACO Algorthm Image Segmetato for Optmal Thresholdg Abstract Usg The ACO Algorthm Image Segmetato for Optmal Thresholdg 陳香伶財務金融系 Despte the fact that the problem of thresholdg has bee qute

More information

Process Quality Evaluation based on Maximum Entropy Principle. Yuhong Wang, Chuanliang Zhang, Wei Dai a and Yu Zhao

Process Quality Evaluation based on Maximum Entropy Principle. Yuhong Wang, Chuanliang Zhang, Wei Dai a and Yu Zhao Appled Mechacs ad Materals Submtted: 204-08-26 ISSN: 662-7482, Vols. 668-669, pp 625-628 Accepted: 204-09-02 do:0.4028/www.scetfc.et/amm.668-669.625 Ole: 204-0-08 204 Tras Tech Publcatos, Swtzerlad Process

More information

Auto-Scalability in Cloud: A Surveyof Energy and Sla Efficient Virtual Machine Consolidation

Auto-Scalability in Cloud: A Surveyof Energy and Sla Efficient Virtual Machine Consolidation SSRG Iteratoal Joural of Computer Scece ad Egeerg (SSRG-IJCSE volume 3 Issue November 06 Auto-Scalablty Cloud: A Surveyof Eergy ad Sla Effcet Vrtual Mache Cosoldato A.Rchard Wllam, Dr.J.Sethlkumar Asst.

More information

Content-Based Image Retrieval Using Associative Memories

Content-Based Image Retrieval Using Associative Memories Proceedgs of the 6th WSEAS It. Coferece o ELECOMMUNICAIONS ad INFORMAICS, Dallas, exas, USA, March 22-24, 2007 99 Cotet-Based Image Retreval Usg Assocatve Memores ARUN KULKARNI Computer Scece Departmet

More information

TDT-2004: ADAPTIVE TOPIC TRACKING AT MARYLAND

TDT-2004: ADAPTIVE TOPIC TRACKING AT MARYLAND TDT-2004: ADAPTIVE TOPIC TRACKING AT MARYLAND Tamer Elsayed, Douglas W. Oard, Davd Doerma Isttute for Advaced r Studes Uversty of Marylad, College Park, MD 20742 Cotact author: telsayed@cs.umd.edu Gary

More information

MINIMIZATION OF THE VALUE OF DAVIES-BOULDIN INDEX

MINIMIZATION OF THE VALUE OF DAVIES-BOULDIN INDEX MIIMIZATIO OF THE VALUE OF DAVIES-BOULDI IDEX ISMO ÄRÄIE ad PASI FRÄTI Departmet of Computer Scece, Uversty of Joesuu Box, FI-800 Joesuu, FILAD ABSTRACT We study the clusterg problem whe usg Daves-Bould

More information

Two step approach for Software Process Control: HLSRGM

Two step approach for Software Process Control: HLSRGM Iteratoal Joural of Emergg Treds & Techology Computer Scece (IJETTCS Web Ste: wwwjettcsorg Emal: edtor@jettcsorg, edtorjettcs@gmalcom Volume, Issue 4, July August 03 ISS 78-686 Two step approach for Software

More information

An Ensemble Approach to Classifier Construction based on Bootstrap Aggregation

An Ensemble Approach to Classifier Construction based on Bootstrap Aggregation A Esemble Approach to Classfer Costructo based o Bootstrap Aggregato Dewa Md. Fard Jahagragar Uversty Dhaka-342, Bagladesh Mohammad Zahdur Rahma Jahagragar Uversty Dhaka-342, Bagladesh Chowdhury Mofzur

More information

On a Sufficient and Necessary Condition for Graph Coloring

On a Sufficient and Necessary Condition for Graph Coloring Ope Joural of Dscrete Matheatcs, 04, 4, -5 Publshed Ole Jauary 04 (http://wwwscrporg/joural/ojd) http://dxdoorg/0436/ojd04400 O a Suffcet ad Necessary Codto for raph Colorg Maodog Ye Departet of Matheatcs,

More information

Interactive Change Detection Using High Resolution Remote Sensing Images Based on Active Learning with Gaussian Processes

Interactive Change Detection Using High Resolution Remote Sensing Images Based on Active Learning with Gaussian Processes Iteractve Chage Detecto Usg Hgh Resoluto Remote Sesg Images Based o Actve Learg wth Gaussa Processes Hu Ru a, Hua Yu a,, Pgpg Huag b, We Yag a a School of Electroc Iformato, Wuha Uversty, 43007 Wuha, Cha

More information

Text Categorization Based on a Similarity Approach

Text Categorization Based on a Similarity Approach Text Categorzato Based o a Smlarty Approach Cha Yag Ju We School of Computer Scece & Egeerg, Uversty of Electroc Scece ad Techology of Cha, Chegdu 60054, P.R. Cha Abstract Text classfcato ca effcetly ehace

More information

Keywords- clustering; naïve Bayesian classifier; boosting; hybrid classifier.

Keywords- clustering; naïve Bayesian classifier; boosting; hybrid classifier. World of Computer Scece ad Iformato Techology Joural (WCSIT) ISSN: 2221-0741 Vol. 1, No. 3, 105-109, 2011. A Hybrd Classfer usg Boostg, Clusterg, ad Naïve Bayesa Classfer A. J. M. Abu Afza, Dewa Md. Fard,

More information

Moving Foreground Detection Based On Spatio-temporal Saliency

Moving Foreground Detection Based On Spatio-temporal Saliency IJCSI Iteratoal Joural of Computer Scece Issues Vol. 10 Issue 1 No 3 Jauary 013 ISSN (Prt): 1694-0784 ISSN (Ole): 1694-0814 www.ijcsi.org 79 Movg Foregroud Detecto Based O Spato-temporal Salecy Yag Xa

More information

Unsupervised Pattern Classification for Categorical Data: A Two Stage Fuzzy Clustering Approach

Unsupervised Pattern Classification for Categorical Data: A Two Stage Fuzzy Clustering Approach 5 Usupervsed Patter Classfcato for Categorcal Data: A Two Stage Fuzzy Clusterg Approach Idrat Saha*, Arba Muhopadhyay, ad Uwal Maul Abstract Clusterg s a popular exploratory patter classfcato tool that

More information

A Double-Window-based Classification Algorithm for Concept Drifting Data Streams

A Double-Window-based Classification Algorithm for Concept Drifting Data Streams 00 IEEE Iteratoal Coferece o Graular Computg A Double-Wdow-based Classfcato Algorthm for Cocept Drftg Data Streams Qu Zhu, Xuegag Hu, Yuhog Zhag, Pepe L, Xdog Wu, School of Computer Scece ad Iformato Egeerg,

More information

Supplementary Information

Supplementary Information Supplemetary Iformato A Self-Trag Subspace Clusterg Algorthm uder Low-Rak Represetato for Cacer Classfcato o Gee Expresso Data Chu-Qu Xa 1, Ke Ha 1, Yog Q 1, Yag Zhag 2, ad Dog-Ju Yu 1,2, 1 School of Computer

More information

Cubic fuzzy H-ideals in BF-Algebras

Cubic fuzzy H-ideals in BF-Algebras OSR Joural of Mathematcs (OSR-JM) e-ssn: 78-578 p-ssn: 39-765X Volume ssue 5 Ver (Sep - Oct06) PP 9-96 wwwosrjouralsorg Cubc fuzzy H-deals F-lgebras Satyaarayaa Esraa Mohammed Waas ad U du Madhav 3 Departmet

More information

ITEM ToolKit Technical Support Notes

ITEM ToolKit Technical Support Notes ITEM ToolKt Notes Fault Tree Mathematcs Revew, Ic. 2875 Mchelle Drve Sute 300 Irve, CA 92606 Phoe: +1.240.297.4442 Fax: +1.240.297.4429 http://www.itemsoft.com Page 1 of 15 6/1/2016 Copyrght, Ic., All

More information

Research Article Fault Diagnosis in Condition of Sample Type Incompleteness Using Support Vector Data Description

Research Article Fault Diagnosis in Condition of Sample Type Incompleteness Using Support Vector Data Description Mathematcal Problems Egeerg Volume 5, Artcle ID 465, pages http://dx.do.org/.55/5/465 Research Artcle Fault Dagoss Codto of Sample Type Icompleteess Usg Support Vector Data Descrpto Hu Y, Zehu Mao,, B

More information

A non-extensive entropy feature and its application to texture classification

A non-extensive entropy feature and its application to texture classification Ths s the author s verso ( post-prt for the publshed artcle to be cted as: Susa, Seba, ad Madasu Hamadlu. "A o-extesve etropy feature ad ts applcato to texture classfcato." Neurocomputg 10 (013: 14-5.

More information

Advanced Information Technology of Slot-Switching Network Schemes for on All-Optical Variable-Length Packet

Advanced Information Technology of Slot-Switching Network Schemes for on All-Optical Variable-Length Packet Joural of Computer Scece 6 (): -, 200 ISS 549-3636 200 Scece Publcatos Advaced Iformato Techology of Slot-Swtchg etwork Schemes for o All-Optcal Varable-Legth Packet Soug Yue Lew, 2 Edward Sek Kh Wog ad

More information

Statistical Techniques Employed in Atmospheric Sampling

Statistical Techniques Employed in Atmospheric Sampling Appedx A Statstcal Techques Employed Atmospherc Samplg A.1 Itroducto Proper use of statstcs ad statstcal techques s ecessary for assessg the qualty of ambet ar samplg data. For a comprehesve dscusso of

More information

Hybrid Parameter Optimization Approach with Adaptive Neuro Fuzzy Inference System for the Software Maintainability

Hybrid Parameter Optimization Approach with Adaptive Neuro Fuzzy Inference System for the Software Maintainability Hbrd Parameter Optmzato Approach wth Adaptve Neuro Fuzz Iferece Sstem for the Software Mataablt P.R Therasa Computer Ceter A.C. Tech, Aa Uverst Chea, Ida therasa.peter@gmal.com P.Vvekaada Dept. of Chemcal

More information

Applying Support Vector Machines to Imbalanced Datasets

Applying Support Vector Machines to Imbalanced Datasets Applyg Support Vector Maches to Imbalaced Datasets Reha Akba 1, Stephe Kwek 1, ad Nathale Japkowcz 2 1 Departmet of Computer Scece, Uversty of Texas at Sa Atoo 6900 N. Loop 1604 W, Sa Atoo, Texas, 78249,

More information

From Math to Efficient Hardware. James C. Hoe Department of ECE Carnegie Mellon University

From Math to Efficient Hardware. James C. Hoe Department of ECE Carnegie Mellon University FFT Compler: From Math to Effcet Hardware James C. Hoe Departmet of ECE Carege Mello Uversty jot wor wth Peter A. Mlder, Fraz Frachett, ad Marus Pueschel the SPIRAL project wth support from NSF ACR-3493,

More information

An Improved Fuzzy C-Means Clustering Algorithm Based on Potential Field

An Improved Fuzzy C-Means Clustering Algorithm Based on Potential Field 07 d Iteratoal Coferece o Advaces Maagemet Egeerg ad Iformato Techology (AMEIT 07) ISBN: 978--60595-457-8 A Improved Fuzzy C-Meas Clusterg Algorthm Based o Potetal Feld Yua-hag HAO, Zhu-chao YU *, X GAO

More information

Analysis of Students' Performance by Using Different Data Mining Classifiers

Analysis of Students' Performance by Using Different Data Mining Classifiers I.J. Moder Educato ad Computer Scece, 2017, 8, 9-15 Publshed Ole August 2017 MECS (http://www.mecs-press.org/) DOI: 10.5815/jmecs.2017.08.02 Aalyss of Studets' Performace by Usg Dfferet Data Mg Classfers

More information

Enumerating XML Data for Dynamic Updating

Enumerating XML Data for Dynamic Updating Eumeratg XML Data for Dyamc Updatg Lau Ho Kt ad Vcet Ng Departmet of Computg, The Hog Kog Polytechc Uversty, Hug Hom, Kowloo, Hog Kog cstyg@comp.polyu.edu.h Abstract I ths paper, a ew mappg model, called

More information

Differentiated Service of Streaming Media Playback Technology

Differentiated Service of Streaming Media Playback Technology Iteratoal Coferece o Advaced Iformato ad Commucato Techology for Educato (ICAICTE 2013) Dfferetated Servce of Streamg Meda Playback Techology CHENG Z-ao 1 MENG Bo 1 WANG Da-hua 1 ZHAO Yue 1 1 Iformatzato

More information

Perceptual Factor Analysis for Speech Enhancement

Perceptual Factor Analysis for Speech Enhancement Perceptual Factor Aalss for Speech Ehacemet Chua-e g ad Je-ug Che Departmet of Computer Scece ad Iformato Egeerg, atoal Cheg ug Uvet, aa, awa, ROC {motor,che}@che.cse.cu.edu.tw Abstract hs paper presets

More information

Advertisement Click-Through Rate Prediction using Multiple Criteria Linear Programming Regression Model

Advertisement Click-Through Rate Prediction using Multiple Criteria Linear Programming Regression Model Avalable ole at www.scecedrect.com Proceda Computer Scece 17 (2013 ) 803 811 Iformato Techology ad Quattatve Maagemet (ITQM2013) Advertsemet Clck-Through Rate Predcto usg Multple Crtera Lear Programmg

More information