Scheme:weka.classifiers.functions.MultilayerPerceptron -L 0.3 -M 0.2 -N 500 -V 0 -S E 20 -H a

Similar documents
IEE 520 Data Mining. Project Report. Shilpa Madhavan Shinde

Why MultiLayer Perceptron/Neural Network? Objective: Attributes:

S2 Text. Instructions to replicate classification results.

Classification Part 4

More on Learning. Neural Nets Support Vectors Machines Unsupervised Learning (Clustering) K-Means Expectation-Maximization

Evaluation Measures. Sebastian Pölsterl. April 28, Computer Aided Medical Procedures Technische Universität München

DATA MINING LAB MANUAL

Evaluating Classifiers

IMPLEMENTATION OF CLASSIFICATION ALGORITHMS USING WEKA NAÏVE BAYES CLASSIFIER

CS145: INTRODUCTION TO DATA MINING

Evaluating Classifiers

The Explorer. chapter Getting started

CS4491/CS 7265 BIG DATA ANALYTICS

Weka ( )

PROJECT 1 DATA ANALYSIS (KR-VS-KP)

A Study of Random Forest Algorithm with implemetation using Weka

Evaluating Machine Learning Methods: Part 1

Machine Learning and Bioinformatics 機器學習與生物資訊學

List of Exercises: Data Mining 1 December 12th, 2015

Evaluating Machine-Learning Methods. Goals for the lecture

CS570: Introduction to Data Mining

Contents. ACE Presentation. Comparison with existing frameworks. Technical aspects. ACE 2.0 and future work. 24 October 2009 ACE 2

Study on Classifiers using Genetic Algorithm and Class based Rules Generation

Data Mining Classification: Alternative Techniques. Imbalanced Class Problem

CS249: ADVANCED DATA MINING

A Comparison of Decision Tree Algorithms For UCI Repository Classification

Best First and Greedy Search Based CFS and Naïve Bayes Algorithms for Hepatitis Diagnosis

Naïve Bayes Classification. Material borrowed from Jonathan Huang and I. H. Witten s and E. Frank s Data Mining and Jeremy Wyatt and others

Data Mining. Practical Machine Learning Tools and Techniques. Slides for Chapter 5 of Data Mining by I. H. Witten, E. Frank and M. A.

Evaluating Classifiers

IV B. Tech I semester (JNTUH-R13)

INTRODUCTION TO MACHINE LEARNING. Measuring model performance or error

Boolean Classification

Data Mining: Classifier Evaluation. CSCI-B490 Seminar in Computer Science (Data Mining)

.. Cal Poly CSC 466: Knowledge Discovery from Data Alexander Dekhtyar.. for each element of the dataset we are given its class label.

Different Classification Technique for Data mining in Insurance Industry using weka

Application of the pessimistic pruning to increase the accuracy of C4.5 algorithm in diagnosing chronic kidney disease

Network Traffic Measurements and Analysis

DATA MINING LECTURE 11. Classification Basic Concepts Decision Trees Evaluation Nearest-Neighbor Classifier

A Comparative Study of Selected Classification Algorithms of Data Mining

Machine learning in fmri

Machine Learning Techniques for Data Mining

MEASURING CLASSIFIER PERFORMANCE

Metrics for Performance Evaluation How to evaluate the performance of a model? Methods for Performance Evaluation How to obtain reliable estimates?

SWETHA ENGINEERING COLLEGE (Approved by AICTE, New Delhi, Affiliated to JNTUA) DATA MINING USING WEKA

Data Mining and Knowledge Discovery: Practice Notes

DATA MINING LECTURE 9. Classification Decision Trees Evaluation

DATA MINING INTRODUCTION TO CLASSIFICATION USING LINEAR CLASSIFIERS

What is Learning? CS 343: Artificial Intelligence Machine Learning. Raymond J. Mooney. Problem Solving / Planning / Control.

ECE 5470 Classification, Machine Learning, and Neural Network Review

Data Mining Concepts & Techniques

CREDIT RISK MODELING IN R. Finding the right cut-off: the strategy curve

2. On classification and related tasks

Classification. Instructor: Wei Ding

NMLRG #4 meeting in Berlin. Mobile network state characterization and prediction. P.Demestichas (1), S. Vassaki (2,3), A.Georgakopoulos (2,3)

CSE Data Mining Concepts and Techniques STATISTICAL METHODS (REGRESSION) Professor- Anita Wasilewska. Team 13

INTRODUCTION TO DATA MINING. Daniel Rodríguez, University of Alcalá

DESIGN AND EVALUATION OF MACHINE LEARNING MODELS WITH STATISTICAL FEATURES

Interpretation and evaluation

Machine Learning for. Artem Lind & Aleskandr Tkachenko

Extreme Learning Machines. Tony Oakden ANU AI Masters Project (early Presentation) 4/8/2014

CISC 4631 Data Mining

Basic Tokenizing, Indexing, and Implementation of Vector-Space Retrieval

Information Management course

Contents Machine Learning concepts 4 Learning Algorithm 4 Predictive Model (Model) 4 Model, Classification 4 Model, Regression 4 Representation

ECLT 5810 Evaluation of Classification Quality

Artificial Neural Networks (Feedforward Nets)

Sergiu Nedevschi Computer Science Department Technical University of Cluj-Napoca

A Comparative Study of Locality Preserving Projection and Principle Component Analysis on Classification Performance Using Logistic Regression

GRIND Gene Regulation INference using Dual thresholding. Ir. W.P.A. Ligtenberg Eindhoven University of Technology

Cross- Valida+on & ROC curve. Anna Helena Reali Costa PCS 5024

DATA MINING LECTURE 9. Classification Basic Concepts Decision Trees Evaluation

Decision trees. For this lab, we will use the Carseats data set from the ISLR package. (install and) load the package with the data set

Evaluation Metrics. (Classifiers) CS229 Section Anand Avati

Data Mining and Knowledge Discovery Practice notes 2

Pattern recognition (4)

Data Mining and Knowledge Discovery: Practice Notes

Classification and Regression

Machine Learning Methods. Majid Masso, PhD Bioinformatics and Computational Biology George Mason University

Functions and Families

โครงการอบรมเช งปฏ บ ต การ น กว ทยาการข อม ลภาคร ฐ (Government Data Scientist)

Comparative Study of J48, Naive Bayes and One-R Classification Technique for Credit Card Fraud Detection using WEKA

Advertisement Image Detection

NETWORK FAULT DETECTION - A CASE FOR DATA MINING

International Journal of Computer Engineering and Applications, Volume XI, Issue XII, Dec. 17, ISSN

Data Mining D E C I S I O N T R E E. Matteo Golfarelli

Advanced Video Content Analysis and Video Compression (5LSH0), Module 8B

Data Mining Classification: Bayesian Decision Theory

DATA MINING AND MACHINE LEARNING. Lecture 6: Data preprocessing and model selection Lecturer: Simone Scardapane

CS 584 Data Mining. Classification 3

Nearest Neighbor 3D Segmentation with Context Features

Machine Learning Software ROOT/TMVA

Probabilistic Classifiers DWML, /27

Use of Synthetic Data in Testing Administrative Records Systems

Dr. Prof. El-Bahlul Emhemed Fgee Supervisor, Computer Department, Libyan Academy, Libya

A Critical Study of Selected Classification Algorithms for Liver Disease Diagnosis

DEEP LEARNING OF COMPRESSED SENSING OPERATORS WITH STRUCTURAL SIMILARITY (SSIM) LOSS

Math 1314 Test 3 Review Material covered is from Lessons 9 15

CLASSIFICATION WITH RADIAL BASIS AND PROBABILISTIC NEURAL NETWORKS

Evaluation of different biological data and computational classification methods for use in protein interaction prediction.

Transcription:

- Rede Neural 1 === Run information === Scheme:weka.classifiers.functions.MultilayerPerceptron -L 0.3 -M 0.2 -N 500 -V 0 -S 12345 -E 20 -H a Relation: Aula9_RedesNeurais-2 Instances: 91 Attributes: 7 CLASS_Y LS GA Rep_EST Rep_PC EST_CUST FORN_VEN Test mode:10-fold cross-validation === Classifier model (full training set) === Sigmoid Node 0 Threshold -6.051034847158921 Node 2 7.231875918969311 Node 3 3.3991329353317155 Node 4 4.947702519335574 Node 5 2.5805109990468535 Sigmoid Node 1 Threshold 6.051369586119442 Node 2-7.2319432700052575

Node 3-3.3993138220794803 Node 4-4.946855344455453 Node 5-2.5809076643611037 Sigmoid Node 2 Threshold 0.46227491349887145 Attrib LS -3.9109792001066492 Attrib GA -6.034103864734481 Attrib Rep_EST 11.646862662729601 Attrib Rep_PC 3.3166713815322164 Attrib EST_CUST 0.48372397773453946 Attrib FORN_VEN 1.3399333860973714 Sigmoid Node 3 Threshold 2.281760521017178 Attrib LS -3.2782185280119145 Attrib GA 4.836058574384419 Attrib Rep_EST 2.5164277863003415 Attrib Rep_PC -4.997717036702611 Attrib EST_CUST -1.3063927679987426 Attrib FORN_VEN -1.046677760591017 Sigmoid Node 4 Threshold -1.406877679743357 Attrib LS -3.3318108024699695 Attrib GA 7.006328212631873 Attrib Rep_EST -5.520361661048657 Attrib Rep_PC 2.617139415485489 Attrib EST_CUST 1.9768519013240589 Attrib FORN_VEN 2.058300384939363 Sigmoid Node 5

Threshold 0.04664594486008929 Attrib LS -2.73260116712376 Attrib GA 1.349773645399601 Attrib Rep_EST -3.3786696624841266 Attrib Rep_PC 3.548082809219273 Attrib EST_CUST 0.3605636586080972 Attrib FORN_VEN 0.43936199473515636 Class I Node 0 Class A Node 1 Time taken to build model: 0.22 seconds === Stratified cross-validation === === Summary === Correctly Classified Instances 78 85.7143 % Incorrectly Classified Instances 13 14.2857 % Kappa statistic 0.715 Mean absolute error 0.1625 Root mean squared error 0.3193 Relative absolute error 32.6796 % Root relative squared error 64.0089 % Total Number of Instances 91 === Detailed Accuracy By Class ===

TP Rate FP Rate Precision Recall F-Measure ROC Area Class 0.816 0.095 0.909 0.816 0.86 0.918 I 0.905 0.184 0.809 0.905 0.854 0.918 A Weighted Avg. 0.857 0.136 0.863 0.857 0.857 0.918 === Confusion Matrix === a b <-- classified as 40 9 a = I 4 38 b = A - Rede Neural 2 === Run information === Scheme:weka.classifiers.functions.MultilayerPerceptron -L 0.9 -M 0.2 -N 500 -V 0 -S 12345 -E 20 -H a Relation: Aula9_RedesNeurais-2 Instances: 91 Attributes: 7 CLASS_Y LS GA Rep_EST Rep_PC EST_CUST FORN_VEN Test mode:10-fold cross-validation

=== Classifier model (full training set) === Sigmoid Node 0 Threshold -2.8374442358427743 Node 2 9.129498752323117 Node 3 1.6743531555904785 Node 4 0.840218962193647 Node 5 7.644058613957676 Sigmoid Node 1 Threshold 2.837442089019373 Node 2-9.12941878696117 Node 3-1.6743420968596612 Node 4-0.8490278179713482 Node 5-7.643776450979976 Sigmoid Node 2 Threshold 0.9083039314575134 Attrib LS -5.36699912477193 Attrib GA -11.462391077743144 Attrib Rep_EST 19.810747642924845 Attrib Rep_PC 5.350044572617745 Attrib EST_CUST 0.7077742445119104 Attrib FORN_VEN 2.629580366317645 Sigmoid Node 3 Threshold 1.9381249648794507 Attrib LS -4.310412454027452 Attrib GA 11.611403905948665

Attrib Rep_EST -2.3121636395038268 Attrib Rep_PC 0.6555443525329911 Attrib EST_CUST -1.514642407347438 Attrib FORN_VEN -1.3069400503703874 Sigmoid Node 4 Threshold -1.3768355408965778 Attrib LS -1.4368325692300836 Attrib GA 4.243843090704246 Attrib Rep_EST -0.27914178687966373 Attrib Rep_PC 2.4214014545647116 Attrib EST_CUST 1.7654084714296145 Attrib FORN_VEN 1.8265192004060684 Sigmoid Node 5 Threshold -3.1161397107798274 Attrib LS -4.931972067010444 Attrib GA 14.613717186213472 Attrib Rep_EST -12.46423983619535 Attrib Rep_PC 4.948227585154165 Attrib EST_CUST 3.3668014254531555 Attrib FORN_VEN 3.392390711374362 Class I Node 0 Class A Node 1 Time taken to build model: 0.14 seconds

=== Stratified cross-validation === === Summary === Correctly Classified Instances 80 87.9121 % Incorrectly Classified Instances 11 12.0879 % Kappa statistic 0.7597 Mean absolute error 0.168 Root mean squared error 0.325 Relative absolute error 33.7858 % Root relative squared error 65.1598 % Total Number of Instances 91 === Detailed Accuracy By Class === TP Rate FP Rate Precision Recall F-Measure ROC Area Class 0.816 0.048 0.952 0.816 0.879 0.874 I 0.952 0.184 0.816 0.952 0.879 0.874 A Weighted Avg. 0.879 0.11 0.89 0.879 0.879 0.874 === Confusion Matrix === a b <-- classified as 40 9 a = I 2 40 b = A - Rede Neural 3 === Run information ===

Scheme:weka.classifiers.functions.MultilayerPerceptron -L 0.2 -M 0.2 -N 5000 -V 0 -S 12345 -E 20 -H a Relation: Aula9_RedesNeurais-2 Instances: 91 Attributes: 7 CLASS_Y LS GA Rep_EST Rep_PC EST_CUST FORN_VEN Test mode:10-fold cross-validation === Classifier model (full training set) === Sigmoid Node 0 Threshold -16.191501252448884 Node 2 9.76114870525198 Node 3 12.879421945654679 Node 4 8.424880171652969 Node 5 12.186648463340886 Sigmoid Node 1 Threshold 16.19150444183219 Node 2-9.761151121851293 Node 3-12.879424653073384 Node 4-8.424876846369653 Node 5-12.186650839583058 Sigmoid Node 2

Threshold -0.46773639627834424 Attrib LS -3.276922093142723 Attrib GA -6.3278349524568105 Attrib Rep_EST 12.421145975540242 Attrib Rep_PC 3.4596824402386255 Attrib EST_CUST 1.5392594536988058 Attrib FORN_VEN 2.4328410934309734 Sigmoid Node 3 Threshold 9.903702371245757 Attrib LS -3.7998372053682243 Attrib GA 12.274413943553922 Attrib Rep_EST 9.681048860351318 Attrib Rep_PC -0.056015167339107726 Attrib EST_CUST -3.0043273402890733 Attrib FORN_VEN -2.30363790702887 Sigmoid Node 4 Threshold -1.7486867995026683 Attrib LS -2.5577042822572165 Attrib GA 10.157206071457052 Attrib Rep_EST -4.803546639388251 Attrib Rep_PC 2.617339857074429 Attrib EST_CUST 2.3669423523220265 Attrib FORN_VEN 2.457250389761146 Sigmoid Node 5 Threshold 5.9841658551857515 Attrib LS 5.839315203123761 Attrib GA -10.571066860443665

Attrib Rep_EST -3.293067780469836 Attrib Rep_PC 29.03484913567614 Attrib EST_CUST -4.879372106255956 Attrib FORN_VEN -4.713811299457585 Class I Node 0 Class A Node 1 Time taken to build model: 1.39 seconds === Stratified cross-validation === === Summary === Correctly Classified Instances 81 89.011 % Incorrectly Classified Instances 10 10.989 % Kappa statistic 0.7804 Mean absolute error 0.1168 Root mean squared error 0.2912 Relative absolute error 23.4877 % Root relative squared error 58.3776 % Total Number of Instances 91 === Detailed Accuracy By Class === TP Rate FP Rate Precision Recall F-Measure ROC Area Class 0.857 0.071 0.933 0.857 0.894 0.939 I 0.929 0.143 0.848 0.929 0.886 0.939 A

Weighted Avg. 0.89 0.104 0.894 0.89 0.89 0.939 === Confusion Matrix === a b <-- classified as 42 7 a = I 3 39 b = A Análise: O modelo 3 apresenta melhor resultado, pois tem o maior acerto (89,01%) e o menor erro relativo médio (23,49%)

2024 © technodocbox.com Privacy Policy | Terms of Service | Feedback