Supervised Descriptive Pattern Mining ， Sebastián

文件名称: Supervised Descriptive Pattern Mining

所属分类: 机器学习

开发工具:

文件大小: 4mb

下载次数: 0

上传时间: 2019-07-04

提供者: weixin_********

下载 (4mb)

不能下载？报告错误

详细说明：Supervised Descriptive Pattern Mining ， Sebastián VenturaJosé María Luna，2018Sebastian ventura Jose maria luna Supervised descriptive Pattern Mining Springer Sebastian ventura Jose maria luna Computer science Computer Science diversity of Cordoba University of cordoba Cordoba, Spain Cordoba. Spain ISBN978-3-31998139-0 ISBN978-3-31998140-6( eBook) https://doi.org/10.10077978-3-319-98140-6 Library of Congress Control Number: 2018953781 o Springer Nature Switzerland AG 2018 This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed The use of general descriptive names, registered names, trademarks, service marks, etc in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, express or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations This Springer imprint is published by the registered company Springer Nature Switzerland AG The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland To Marta and Laura, my favorite noise makers S. Ventura To my son Luca, the one who makes me smile every aay. J.M. Luna Acknowledgments The authors would like to thank the Springer editorial team for giving them the opportunity to publish this book, for their great support toward the preparation and completion of this work, and their valuable editing suggestions to improve the organization and readability of the manuscript The authors also want to thank anyone who helped them during the preparation of the book, whose comments were very helpful for improving its quality The authors would also like to thank the Spanish Ministry of Economy and the European Fund for Regional Development for supporting their research toward the project TIN2017-83445-P Contents 1 Introduction to Supervised Descriptive Pattern Mining 1.1 Patterns in Data Analysis 1.2 Pattern Mining: Types of Patterns and advanced data Types .... 3 1. 2.1 Frequent/Infrequent Patterns 1. 2. 2 Positive/Negative patterns 1. 2.3 Maximal/ Closed/ Colossal patterns 7 1. 2. 4 Condensed Patterns 9 1.2.5 Patterns on Advanced Data Types 1.3 Supervised Descriptive Patterns .................... 15 1. 3.1 Contrast Sets 15 1.3.2 Emerging Patterns 16 1.3.3 Subgroup discovery.…… 17 1. 3. 4 Class association rules 18 1.3.5 Exceptional models 20 1.3.6 Other Forms of Supervised Descriptive Patterns....... 21 1. 4 Scalability Issues 2 1.4. I Heuristic Approaches..…… 1. 4.2 New Data Structures 23 1.4.3 Parallel Computin 24 1. 4.4 MapReduce Framework 25 Refe 26 2 Contrast sets 33 2.1 Introduction 2.2 Task definition 2.2.1 Quality Measures 2.2.2 Tree Structures ......................... 38 Contents 2.3 Algorithms for Mining Contrast Sets................. 43 2.3.1 STUCCO 43 2.3.2 CIGAR 46 2.3.3 CSM-SD 48 2.3.4 Additional approaches..................... 49 References 50 3 Emerging Patterns.............................. 53 3.1 Introduction 53 3.2 Task definition 3.2.1 Problem Decomposition…… 3.2.2 Types of Emerging Patterns.…………,57 3.3 Algorithms for Mining Emerging Patterns ............... 58 3.3.1 Border-Based algorithms 58 3.3.2 Constraint-Based Algorithms................. 61 3.3.3 Tree-Based Algorithms 64 3.3.4 Evolutionary Fuzzy System-Based Algorithms 67 References. ....................................................................69 4 Subgroup Discovery........................ 4.1 Introduction 71 4.2 Task Definition 4.2.1 Quality measures 73 4.2.2 Unifying Related Tasks............. 77 4.3 Algorithms for Subgroup Discovery............ 78 4.3.1 Extensions of Classification Algorithms 4.3.2 Extensions of Association Rule Mining Algorithms 81 4.3.3 Evolutionary Algorithms........... 84 4.3.4 Big Data Approaches 93 References 96 5 Class association rules 5.1 Introducti 5.2 Task Definition 101 5.2.1 Quality me 5.2.2 Class Association Rules for Descriptive analysis 105 5.2.3 Class Association Rules for Predictive Analysis. .............106 5.2.4 Related Tasks 5.3 Algorithms for Class Association Rules. ..............................107 5.3.1 Algorithms for Descriptive Analysis 107 5.3.2 Algorithms for Predictive Analysis 119 References 126 Contents 6 Exceptional Models 129 6.1 Introduction 129 6.2 Task definition ...131 6.2.1 Original Model Classes 132 6.2.2 Rank Correlation Model Class ................. 135 6.2.3 Related Tasks ..................................................138 6.3 Algorithms for Mining Exceptional models 139 6.3.1 Exceptional Model Mining.……….139 6.3.2 Exceptional Preference Mining.……,142 6.3.3 Exceptional Relationship Mining 144 References 148 7 Other Forms of Supervised Descriptive Pattern Mining........ 151 7.1 Introd 151 7.2 Additional tasks .153 7.2.1 Closed Sets for Labeled data 153 7.2.2 Bump Hunting 156 7.2.3 Impact Rules .......................... 162 7.2.4 Discrimination Discovery .164 7.2.5 Describing Black Box Models 166 7.2.6 Change Mining 167 Refe 8 Successful Applications.……,…,…,……,171 8.1 Introduction 171 8.2 Applications∴……173 8.2.1 Medicine 8. 2. 2 Education 175 8. 2. 3 Sociol 178 8.2.4E y and ener 179 8.2.5 Other Problems . ...............................................182 Referene 183 Chapter 1 Introduction to Supervised Descriptive ck 1 Pattern Mining abstract This chapter introduces the supervised descriptive pattern mining task to the reader, providing him/her with the concept of patterns as well as presenting a description of the type of patterns usually found in literature Patterns on advanced data types are also defined, denoting the usefulness of sequential and spatiotemporal patterns, patterns on graphs, high utility patterns, uncertain patterns, along with patterns defined on multiple-instance domains. The utility of the supervised descrip tive pattern mining task is analysed and its main subtasks are formally described including contrast sets, emerging patterns, subgroup discovery, class association rules, exceptional models, among others. Finally, the importance of analysing the computational complexity in the pattern mining field is also considered, examining different ways of reducing this complexity. 1.1 Patterns in Data Analysis In many application fields, there is a real incentive to collect, manage and transform w data into significant and meaningful information that may be used for a subsequent analysis that leads better decision making [79]. In these fields there is therefore a growing interest in data analysis, which is concerned with the development of methods and techniques for making sense of data [31]. When talking about data analysis, the key element is the pattern, representing any type of homogeneity and regularity in data and serving as a good descriptor of intrinsic and important properties of data [3]. A pattern is a set of elements(items)that are related in a database. Formally speaking, a pattern P in a database $2 is defined as a subset of items I=i,.,inE 32, i.e. PCI, that describes valuable features of data [83]. Additionally, the length or size of a pattern P is denoted as the number of single items that it comprises, and those patterns consisting of a unique item are called singletons Pattern mining is a really alluring task whose aim is to find all patterns that are present in at least a fraction of the transactions(records) in a database. This task is being applied to more and more application fields nowadays, market basket analysis [12] being the first application domain in which it was correctly applied. In o Springer Nature Switzerland AG 2018 S. Ventura, J. M. Luna, Supervised Descriptive Pattern Mining https://doi.org/10.10077978-3-319-98140-6

(系统自动生成,下载前可以参看下载内容)