Empirical Investigation of Decision Tree Extraction From

Total Page:16

File Type:pdf, Size:1020Kb

Empirical Investigation of Decision Tree Extraction From EMPIRICAL INVESTIGATION OF DECISION TREE EXTRACTION FROM NEURAL NETWORKS A thesis presented to the faculty of the Fritz J. and Dolores H. Russ College of Engineering and Technology of Ohio University In partial fulfillment of the requirements for the degree Master of Science Maimuna H. Rangwala June 2006 This thesis entitled EMPIRICAL INVESTIGATION OF DECISION TREE EXTRACTION FROM NEURAL NETWORKS by MAIMUNA H. RANGWALA has been approved for the Department of Industrial and Manufacturing Systems Engineering and the Russ College of Engineering and Technology by Gary R. Weckman Associate Professor of Industrial & Manufacturing Systems Engineering R. Dennis Irwin Dean, Fritz J. and Dolores H. Russ College of Engineering and Technology ABSTRACT RANGWALA, MAIMUNA H., M.S., June 2006. Industrial and Manufacturing Systems Engineering EMPIRICAL INVESTIGATION OF DECISION TREE EXTRACTION FROM NEURAL NETWORKS (201 pp.) Director of Thesis: Gary R. Weckman The purpose of this thesis is to develop heuristics for employing Trepan, an algorithm for extracting decision trees from neural networks. Typically, several parameters need to be chosen to obtain a satisfactory performance of the algorithm. The current understanding of the various interactions between these is not well understood. By empirically evaluating the performance of the algorithm on a test set of databases chosen from benchmark machine learning and real world problems, several heuristics are proposed to explain and improve the performance of the algorithm. The experimentation is further validated by performance statistic measures. The algorithm is extended to work for multi-class regression problems and its ability to comprehend generalized feedforward networks is investigated. This work thus serves to provide improvements, an increased understanding of the behavior of the algorithm and heuristics to choose parameters for a better performance. Approved: Gary R. Weckman Associate Professor of Industrial and Manufacturing Systems Engineering Dedicated to my Father Prof. H. T. RANGWALA (1942-2006) and my Sister Fatema Rangwala (1988-2006) 5 TABLE OF CONTENTS ABSTRACT........................................................................................................................3 LIST OF TABLES.............................................................................................................. 8 LIST OF FIGURES .......................................................................................................... 11 CHAPTER 1. INTRODUCTION .................................................................................... 13 1.1 MACHINE LEARNING.......................................................................................... 13 1.2 CLASSIFICATION ALGORITHMS .......................................................................... 14 1.3 RESEARCH OBJECTIVES ..................................................................................... 14 1.4 THESIS OVERVIEW ............................................................................................. 16 CHAPTER 2. BACKGROUND AND LITERATURE REVIEW .................................. 17 2.1 ARTIFICIAL NEURAL NETWORKS ....................................................................... 17 2.1.1 Neural Network Architecture........................................................................ 17 2.1.2 Neural Network Training.............................................................................. 24 2.1.3 Neural Networks for Classification and Regression..................................... 25 2.1.4 Rule Extraction from Neural Networks ........................................................ 26 2.2 DECISION TREES ................................................................................................ 27 2.2.1 Decision Tree Classification......................................................................... 27 2.2.2 Decision Tree Applications........................................................................... 29 2.3 C4.5 ALGORITHM .............................................................................................. 30 6 2.3.1 Information Gain, Entropy Measure and Gain Ratio.................................... 31 2.4 TREPAN ALGORITHM....................................................................................... 37 2.4.1 M-of-N Splitting tests ................................................................................... 39 2.4.2 Single Test TREPAN and Disjunctive TREPAN ......................................... 40 CHAPTER 3. METHODOLOGY ................................................................................... 41 3.1 PHASE 1 ............................................................................................................. 43 3.2 PHASE 2 ............................................................................................................. 44 3.2.1 Datasets......................................................................................................... 45 3.2.2 Neural Network Modeling ............................................................................ 56 3.3 PHASE 3 ............................................................................................................. 57 3.4 PHASE 4 ............................................................................................................. 60 3.5 PERFORMANCE MEASURES ................................................................................. 62 3.5.1 Classification Accuracy................................................................................ 62 3.5.2 Comprehensibility......................................................................................... 64 CHAPTER 4. RESULTS AND DISCUSSION............................................................... 65 4.1 INVESTIGATE AND EXTEND TREPAN................................................................ 65 4.2 DATASET ANALYSIS .......................................................................................... 65 4.2.1 Corrosion....................................................................................................... 65 4.2.2 Outages ......................................................................................................... 79 4.2.3 Iris ................................................................................................................. 89 4.2.4 Body Fat........................................................................................................ 94 7 4.2.5 Saginaw Bay............................................................................................... 101 4.2.6 Admissions.................................................................................................. 110 CHAPTER 5. CONCLUSIONS AND FUTURE RESEARCH .................................... 118 5.1 SUMMARY AND DISCUSSION ............................................................................ 118 5.1.1 Accuracy ..................................................................................................... 119 5.1.2 Comprehensibility....................................................................................... 120 5.2 HEURISTICS...................................................................................................... 120 5.3 CONCLUSIONS.................................................................................................. 122 5.4 FUTURE RESEARCH.......................................................................................... 123 REFERENCES ............................................................................................................... 125 APPENDIX A: WEIGHTS AND NETWORK FILE FORMATS (GFF)...................... 131 APPENDIX B: CORROSION RESULTS...................................................................... 134 APPENDIX C: OUTAGES RESULTS.......................................................................... 143 APPENDIX D: IRIS RESULTS..................................................................................... 153 APPENDIX E: BODY FAT RESULTS......................................................................... 157 APPENDIX F: SAGINAW BAY RESULTS................................................................. 161 APPENDIX G: ADMISSIONS RESULTS.................................................................... 178 8 LIST OF TABLES Table 2.1: Activation Functions used in Neural Networks............................................... 21 Table 2.2: Play Tennis Example Dataset .......................................................................... 34 Table 3.1: Dataset Summary............................................................................................. 44 Table 3.2: Iris--Sample Dataset ........................................................................................ 45 Table 3.3: Body Fat -- Sample Dataset............................................................................. 47 Table 3.4: Body Fat Class Labels ..................................................................................... 47 Table 3.5: Saginaw Bay -- Sample Dataset ...................................................................... 50 Table 3.6: Chlorophyll Level Class Labels....................................................................... 50 Table 3.7: Corrosion -- Sample Dataset............................................................................ 52 Table 3.8: Corrosion Class labels ....................................................................................
Recommended publications
  • Black Box Explanation by Learning Image Exemplars in the Latent Feature Space
    Black Box Explanation by Learning Image Exemplars in the Latent Feature Space Riccardo Guidotti1, Anna Monreale2, Stan Matwin3;4, and Dino Pedreschi2 1 ISTI-CNR, Pisa, Italy, [email protected] 2 University of Pisa, Italy, [email protected] 3 Dalhousie University, [email protected] 4 Institute of Computer Scicne, Polish Academy of Sciences Abstract. We present an approach to explain the decisions of black box models for image classification. While using the black box to label im- ages, our explanation method exploits the latent feature space learned through an adversarial autoencoder. The proposed method first gener- ates exemplar images in the latent feature space and learns a decision tree classifier. Then, it selects and decodes exemplars respecting local decision rules. Finally, it visualizes them in a manner that shows to the user how the exemplars can be modified to either stay within their class, or to be- come counter-factuals by \morphing" into another class. Since we focus on black box decision systems for image classification, the explanation obtained from the exemplars also provides a saliency map highlighting the areas of the image that contribute to its classification, and areas of the image that push it into another class. We present the results of an experimental evaluation on three datasets and two black box models. Be- sides providing the most useful and interpretable explanations, we show that the proposed method outperforms existing explainers in terms of fidelity, relevance, coherence, and stability. Keywords: Explainable AI, Adversarial Autoencoder, Image Exemplars. 1 Introduction Automated decision systems based on machine learning techniques are widely used for classification, recognition and prediction tasks.
    [Show full text]
  • On the Boosting Ability of Top-Down Decision Tree Learning Algorithms
    On the Bo osting AbilityofTop-Down Decision Tree Learning Algorithms Michael Kearns Yishay Mansour AT&T Research Tel-Aviv University May 1996 Abstract We analyze the p erformance of top-down algorithms for decision tree learning, such as those employed by the widely used C4.5 and CART software packages. Our main result is a pro of that such algorithms are boosting algorithms. By this we mean that if the functions that lab el the internal no des of the decision tree can weakly approximate the unknown target function, then the top-down algorithms we study will amplify this weak advantage to build a tree achieving any desired level of accuracy. The b ounds we obtain for this ampli catio n showaninteresting dep endence on the splitting criterion used by the top-down algorithm. More precisely, if the functions used to lab el the internal no des have error 1=2 as approximation s to the target function, then for the splitting criteria used by CART and C4.5, trees 2 2 2 O 1= O log 1== of size 1= and 1= resp ectively suce to drive the error b elow .Thus for example, a small constant advantage over random guessing is ampli ed to any larger constant advantage with trees of constant size. For a new splitting criterion suggested by our analysis, the much stronger 2 O 1= b ound of 1= which is p olynomial in 1= is obtained, whichisprovably optimal for decision tree algorithms. The di ering b ounds have a natural explanation in terms of concavity prop erties of the splitting criterion.
    [Show full text]
  • Boosting with Multi-Way Branching in Decision Trees
    Boosting with Multi-Way Branching in Decision Trees Yishay Mansour David McAllester AT&T Labs-Research 180 Park Ave Florham Park NJ 07932 {mansour, dmac }@research.att.com Abstract It is known that decision tree learning can be viewed as a form of boosting. However, existing boosting theorems for decision tree learning allow only binary-branching trees and the generalization to multi-branching trees is not immediate. Practical decision tree al­ gorithms, such as CART and C4.5, implement a trade-off between the number of branches and the improvement in tree quality as measured by an index function. Here we give a boosting justifica­ tion for a particular quantitative trade-off curve. Our main theorem states, in essence, that if we require an improvement proportional to the log of the number of branches then top-down greedy con­ struction of decision trees remains an effective boosting algorithm. 1 Introduction Decision trees have been proved to be a very popular tool in experimental machine learning. Their popularity stems from two basic features - they can be constructed quickly and they seem to achieve low error rates in practice. In some cases the time required for tree growth scales linearly with the sample size. Efficient tree construction allows for very large data sets. On the other hand, although there are known theoretical handicaps of the decision tree representations, it seem that in practice they achieve accuracy which is comparable to other learning paradigms such as neural networks. While decision tree learning algorithms are popular in practice it seems hard to quantify their success ,in a theoretical model.
    [Show full text]
  • Inductive Bias in Decision Tree Learning • Issues in Decision Tree Learning • Summary
    Machine Learning Decision Tree Learning Artificial Intelligence & Computer Vision Lab School of Computer Science and Engineering Seoul National University Overview • Introduction • Decision Tree Representation • Learning Algorithm • Hypothesis Space Search • Inductive Bias in Decision Tree Learning • Issues in Decision Tree Learning • Summary AI & CV Lab, SNU 2 Introduction • Decision tree learning is a method for approximating discrete-valued target function • The learned function is represented by a decision tree • Decision tree can also be re-represented as if-then rules to improve human readability AI & CV Lab, SNU 3 Decision Tree Representation • Decision trees classify instances by sorting them down the tree from the root to some leaf node • A node – Specifies some attribute of an instance to be tested • A branch – Corresponds to one of the possible values for an attribute AI & CV Lab, SNU 4 Decision Tree Representation (cont.) Outlook Sunny Overcast Rain Humidity Yes Wind High Normal Strong Weak No Yes No Yes A Decision Tree for the concept PlayTennis AI & CV Lab, SNU 5 Decision Tree Representation (cont.) • Each path corresponds to a conjunction of attribute Outlook tests. For example, if the instance is (Outlook=sunny, Temperature=Hot, Sunny Rain Humidity=high, Wind=Strong) then the path of Overcast (Outlook=Sunny ∧ Humidity=High) is matched so that the target value would be NO as shown in the tree. Humidity Wind • A decision tree represents a disjunction of Yes conjunction of constraints on the attribute values of instances. For example, three positive instances can High Normal Strong Weak be represented as (Outlook=Sunny ∧ Humidity=normal) ∨ (Outlook=Overcast) ∨ (Outlook=Rain ∧Wind=Weak) as shown in the tree.
    [Show full text]
  • A Prediction for Student's Performance Using Decision Tree ID3 Method
    International Journal of Scientific & Engineering Research, Volume 5, Issue 7, July-2014 1329 ISSN 2229-5518 Data Mining: A prediction for Student's Performance Using Decision Tree ID3 Method D.BHU LAKSHMI, S. ARUNDATHI, DR.JAGADEESH Abstract— Knowledge Discovery and Data Mining (KDD) is a multidisciplinary area focusing upon methodologies for extracting useful knowledge from data and there are several useful KDD tools to extracting the knowledge. This knowledge can be used to increase the quality of education. But educational institution does not use any knowledge discovery process approach on these data. Data mining can be used for decision making in educational system. A decision tree classifier is one of the most widely used supervised learning methods used for data exploration based on divide & conquer technique. This paper discusses use of decision trees in educational data mining. Decision tree algorithms are applied on students’ past performance data to generate the model and this model can be used to predict the students’ performance. The most useful data mining techniques in educational database is classification, the decision tree (ID3) method is used here. Index Terms— Educational Data Mining, Classification, Knowledge Discovery in Database (KDD), ID3 Algorithm. 1. Introduction the students, prediction about students’ The advent of information technology in various performance and so on, the classification task is fields has lead the large volumes of data storage in used to evaluate student’s performance and as various formats like records, files, documents, there are many approaches that are used for data images, sound, videos, scientific data and many classification, the decision tree method is used new data formats.
    [Show full text]
  • Galaxy Classification with Deep Convolutional Neural Networks
    c 2016 Honghui Shi GALAXY CLASSIFICATION WITH DEEP CONVOLUTIONAL NEURAL NETWORKS BY HONGHUI SHI THESIS Submitted in partial fulfillment of the requirements for the degree of Master of Science in Electrical and Computer Engineering in the Graduate College of the University of Illinois at Urbana-Champaign, 2016 Urbana, Illinois Adviser: Professor Thomas S. Huang ABSTRACT Galaxy classification, using digital images captured from sky surveys to de- termine the galaxy morphological classes, is of great interest to astronomy researchers. Conventional methods rely heavily on a few handcrafted mor- phological features while popular feature extraction methods that developed for natural images are not suitable for galaxy images. Deep convolutional neural networks (CNNs) are able to learn powerful features from images by hierarchical convolutional and pooling operations. This work applies state-of- the-art deep CNN technologies to galaxy classification for both a regression task and multi-class classification tasks. We also implement and compare the performance with several different conventional machine learning algorithms for a classification sub-task. Our experiments show that convolutional neural networks are able to learn representative features automatically and achieve high performance, surpassing both human recognition and other machine learning methods. ii To my family, especially my wife, and my friends near or far. To my adviser, to whom I owe much thanks! iii ACKNOWLEDGMENTS I would like to acknowledge my adviser Professor Thomas Huang, who has given me lots of guidance, support, and visionary insights. I would also like to acknowledge Professor Robert Brunner who led me to the topic and granted me lots of help during the research.
    [Show full text]
  • X-Trepan: a Multi Class Regression and Adapted Extraction of Comprehensible Decision Tree in Artificial Neural Networks
    X-TREPAN: A MULTI CLASS REGRESSION AND ADAPTED EXTRACTION OF COMPREHENSIBLE DECISION TREE IN ARTIFICIAL NEURAL NETWORKS Awudu Karim1, Shangbo Zhou2 College of Computer Science, Chongqing University, Chongqing, 400030, China. [email protected] ABSTRACT In this work, the TREPAN algorithm is enhanced and extended for extracting decision trees from neural networks. We empirically evaluated the performance of the algorithm on a set of databases from real world events. This benchmark enhancement was achieved by adapting Single-test TREPAN and C4.5 decision tree induction algorithms to analyze the datasets. The models are then compared with X-TREPAN for comprehensibility and classification accuracy. Furthermore, we validate the experimentations by applying statistical methods. Finally, the modified algorithm is extended to work with multi-class regression problems and the ability to comprehend generalized feed forward networks is achieved. KEYWORDS: Neural Network, Feed Forward, Decision Tree, Extraction, Classification, Comprehensibility. 1. INTRODUCTION Artificial neural networks are modeled based on the human brain architecture. They offer a means of efficiently modeling large and complex problems in which there are hundreds of independent variables that have many interactions. Neural networks generate their own implicit rules by learning from examples. Artificial neural networks have been applied to a variety of problem domains [1] such as medical diagnostics [2], games [3], robotics [4], speech generation [5] and speech recognition [6]. The generalization ability of neural networks has proved to be superior to other learning systems over a wide range of applications [7]. However despite their relative success, the further adoption of neural networks in some areas has been impeded due to their inability to explain, in a comprehensible form, how a decision has been arrived at.
    [Show full text]
  • Evaluation and Comparison of Word Embedding Models, for Efficient Text Classification
    Evaluation and comparison of word embedding models, for efficient text classification Ilias Koutsakis George Tsatsaronis Evangelos Kanoulas University of Amsterdam Elsevier University of Amsterdam Amsterdam, The Netherlands Amsterdam, The Netherlands Amsterdam, The Netherlands [email protected] [email protected] [email protected] Abstract soon. On the contrary, even non-traditional businesses, like 1 Recent research on word embeddings has shown that they banks , start using Natural Language Processing for R&D tend to outperform distributional models, on word similarity and HR purposes. It is no wonder then, that industries try to and analogy detection tasks. However, there is not enough take advantage of new methodologies, to create state of the information on whether or not such embeddings can im- art products, in order to cover their needs and stay ahead of prove text classification of longer pieces of text, e.g. articles. the curve. More specifically, it is not clear yet whether or not theusage This is how the concept of word embeddings became pop- of word embeddings has significant effect on various text ular again in the last few years, especially after the work of classifiers, and what is the performance of word embeddings Mikolov et al [14], that showed that shallow neural networks after being trained in different amounts of dimensions (not can provide word vectors with some amazing geometrical only the standard size = 300). properties. The central idea is that words can be mapped to In this research, we determine that the use of word em- fixed-size vectors of real numbers. Those vectors, should be beddings can create feature vectors that not only provide able to hold semantic context, and this is why they were a formidable baseline, but also outperform traditional, very successful in sentiment analysis, word disambiguation count-based methods (bag of words, tf-idf) for the same and syntactic parsing tasks.
    [Show full text]
  • Decision Tree Classification and Forecasting of Pricing Time Series
    Decision Tree Classification and Forecasting of Pricing Time Series Data EMIL LUNDKVIST Master’s Degree Project Stockholm, Sweden July 2014 TRITA-XR-EE-RT 2014:017 Abstract Many companies today, in different fields of operations and sizes, have access to a vast amount of data which was not available only a couple of years ago. This situation gives rise to questions regarding how to organize and use the data in the best way possible. In this thesis a large database of pricing data for products within various market segments is analysed. The pricing data is from both external and internal sources and is therefore confidential. Because of the confidentiality, the labels from the database are in this thesis substituted with generic ones and the company is not referred to by name, but the analysis is carried out on the real data set. The data is from the beginning unstructured and difficult to overlook. Therefore, it is first classified. This is performed by feeding some manual training data into an algorithm which builds a decision tree. The decision tree is used to divide the rest of the products in the database into classes. Then, for each class, a multivariate time series model is built and each product’s future price within the class can be predicted. In order to interact with the classification and price prediction, a front end is also developed. The results show that the classification algorithm both is fast enough to operate in real time and performs well. The time series analysis shows that it is possible to use the information within each class to do predictions, and a simple vector autoregressive model used to perform it shows good predictive results.
    [Show full text]
  • Chapter 2. Machine Learning Overview
    Machine Learning Overview 2 2.1 Overview Learning denotes the process of acquiring new declarative knowledge, the organization of new knowledge into general yet effective representations, and the discovery of new facts and theories through observation and experimentation. Learning is one of the most important skills that mankind can master, which also renders us different from the other animals on this planet. To provide an example, according to our past experiences, we know the sun rises from the east and falls to the west; the moon rotates around the earth; 1 year has 365 days, which are all knowledge we derive from our past life experiences. As computers become available, mankind has been striving very hard to implant such skills into computers. For the knowledge which are clear for mankind, they can be explicitly represented in program as a set of simple reasoning rules. Meanwhile, in the past couple of decades, an extremely large amount of data is being generated in various areas, including the World Wide Web (WWW), telecommunication, climate, medical science, transportation, etc. For these applications, the knowledge to be detected from such massive data can be very complex that can hardly be represented with any explicit fine-detailed specification about what these patterns are like. Solving such a problem has been, and still remains, one of the most challenging and fascinating long-range goals of machine learning. Machine learning is one of the disciplines, which aims at endowing programs with the ability to learn and adapt. In machine learning, experiences are usually represented as data, and the main objective of machine learning is to derive models from data that can capture the complicated hidden patterns.
    [Show full text]
  • A Commodity Classification Framework Based on Machine
    S S symmetry Article A Commodity Classification Framework Based on Machine Learning for Analysis of Trade Declaration Mingshu He 1 , Xiaojuan Wang 1,*, Chundong Zou 1, Bingying Dai 2 and Lei Jin 3 1 School of Electronic Engineering, Beijing University of Posts and Telecommunications, Beijing 100876, China; [email protected] (M.H.); [email protected] (C.Z.) 2 Department of Statistics, Colorado State University, Fort Collins, CO 80523, USA; [email protected] 3 School of Computer Science, Beijing University of Posts and Telecommunications, Beijing 100876, China; [email protected] * Correspondence: [email protected] Abstract: Text, voice, images and videos can express some intentions and facts in daily life. By understanding these contents, people can identify and analyze some behaviors. This paper focuses on the commodity trade declaration process and identifies the commodity categories based on text information on customs declarations. Although the technology of text recognition is mature in many application fields, there are few studies on the classification and recognition of customs declaration goods. In this paper, we proposed a classification framework based on machine learning (ML) models for commodity trade declaration that reaches a high rate of accuracy. This paper also proposed a symmetrical decision fusion method for this task based on convolutional neural network (CNN) and transformer. The experimental results show that the fusion model can make up for the shortcomings of the two original models and some improvements have been made. In the two datasets used in this Citation: He, M.; Wang, X.; Zou, C.; Dai, B.; Jin, L. A Commodity paper, the accuracy can reach 88% and 99%, respectively.
    [Show full text]
  • Network Classification and Categorization Paper
    Network Classification and Categorization Paper James P. Canning, Emma E. Ingram, Adriana M. Ortiz and Karl R. B. Schmitt June 2017 Abstract Networks are often labeled according to the underlying phenomena that they represent, such as re- tweets, protein interactions, or web page links. Our research seeks to determine if we can use machine learning techniques to gain a better understanding of the categories of networks on the Network Repos- itory (www.networkrepository.com) and then classify the unlabeled networks into categories that make sense. It is generally believed that networks from different categories have inherently unique network characteristics. Our research provides conclusive evidence to validate this belief by presenting the results of global network clustering and classification into common categories using machine learning algorithms. The machine learning techniques of Decisions Trees, Random Forests, Linear Support Vector Classifica- tion and Gaussian Na¨ıve Bayes were applied to a 14-feature `identifying vector' for each graph. During cross-validation, the best technique, Gaussian Na¨ıve Bayes, achieved an accuracy of 92.8%. After train- ing the machine learning algorithm it was applied to a collection of initially unlabeled graphs from the Network Repository (www.networkrepository.com). Results were then manually checked by determining (when possible) original sources for these graphs. We conclude by examining the accuracy of our results and discussing how future researchers can make use of this process. 1 Introduction Network science is a field which seeks to understand the world by representing interactions and relationships between objects as graphs. By organizing data into networks, we can better understand how objects of interest are connected.
    [Show full text]