发帖
雷达卡
by Tom Rutkowski (Author)
About this Book
The book proposes techniques, with an emphasis on the financial sector, which will make recommendation systems both accurate and explainable. The vast majority of AI models work like black box models. However, in many applications, e.g., medical diagnosis or venture capital investment recommendations, it is essential to explain the rationale behind AI systems decisions or recommendations. Therefore, the development of artificial intelligence cannot ignore the need for interpretable, transparent, and explainable models. First, the main idea of the explainable recommenders is outlined within the background of neuro-fuzzy systems. In turn, various novel recommenders are proposed, each characterized by achieving high accuracy with a reasonable number of interpretable fuzzy rules. The main part of the book is devoted to a very challenging problem of stock market recommendations. An original concept of the explainable recommender, based on patterns from previous transactions, is developed; it recommends stocks that fit the strategy of investors, and its recommendations are explainable for investment advisers.
Brief Contents
1 Introduction 1
1.1 The Purpose of This Book 1
1.2 The Pursuit of Explainable Artificial Intelligence 2
1.3 Recommender Systems 8
1.4 Interpretability of Machine Learning Models 10
1.5 The Content andMain Results of the Book 13
References 16
2 Neuro-Fuzzy Approach and Its Application in Recommender Systems 23
2.1 Neuro-Fuzzy Systems as Recommenders 23
2.2 Fuzzy IF-THEN Rules and Learning Ability of the Recommenders 24
2.3 Interpretability and Explainability of Neuro-Fuzzy Recommenders 26
2.4 Rule Generation From Data 29
2.4.1 Input and Output Data for Neuro-Fuzzy Recommenders 30
2.4.2 Wang-Mendel Method of Rule Generation 30
2.4.3 Nozaki-Ishibuchi-Tanaka Method 33
2.5 Fuzzy IF-THEN Rules in Recommendation Problems 35
2.6 Classification in Recommenders 36
2.6.1 Neuro-Fuzzy Classifiers 36
2.6.2 One-Class Classifiers 37
2.6.3 Classification in Content-Based Recommender Systems 38
References 39
3 Novel Explainable Recommenders Based on Neuro-Fuzzy Systems 43
3.1 Recommender A 43
3.1.1 Feature Encoding 43
3.1.2 Description of the Proposed Recommender A 44
3.1.3 Systems Performance Evaluation 48
3.2 Recommender B 50
3.2.1 Introduction to the Proposed Recommender B 50
3.2.2 Description of the Recommender B 51
3.2.3 Criteria ofBalance EvaluationBetween Recommender Accuracy and Interpretability 53
3.2.4 Recommenders Performance Evaluation 54
3.2.5 Interpretability and Explainability of the Recommender 62
3.3 Recommender C 62
3.3.1 Nominal Attribute Values Encoding 63
3.3.2 Various Neuro-Fuzzy Systems as the Proposed Recommender C 66
3.3.3 Illustration of the Recommender Performance 68
3.4 Conclusions Concerning Recommenders A, B, and C 69
References 72
4 Explainable Recommender for Investment Advisers 75
4.1 Introduction to the Real-Life Application of the Proposed Recommender 75
4.2 Statement of the Problem 76
4.3 Description of the Datasets and Feature Selection 79
4.3.1 Data Enrichment and Dataset Preparation 79
4.3.2 Description of Selected Attributes—Simplified Version 79
4.3.3 Multidimensional Data Visualization 82
4.4 Definition of Fuzzy Sets 84
4.5 Fuzzy Rule Generation 94
4.6 Results of the System Performance 97
4.6.1 Recommendations Produced by the Recommender 97
4.6.2 Visualization of the Recommender Results 99
4.6.3 Explanations of the Recommendations 105
4.6.4 Evaluation of the Recommender Performance 111
4.7 Conclusions Concerning the Proposed One-Class Recommender 112
References 118
5 Summary and Final Remarks 121
5.1 Summary of the Contributions and Novelties 121
5.2 Future Research 125
5.3 Author’s Contribution 126
References 127
Appendix A: Description of Attributes—Full Version 129
Appendix B: Fuzzy IF-THEN Rules 133
Appendix C: Fuzzy Rules - Full Version 139
Appendix D: Histograms of Attribute Values 143
Appendix E: Fuzzy Sets for Particular Attributes 153
Appendix F: Fuzzy Sets for Single Data Points 161
Series : Studies in Computational Intelligence, 964
Publisher : Springer; 1st ed. 2021 edition (June 8, 2021)
Language : English
Pages : 186
ISBN-10 : 3030755207
ISBN-13 : 978-3030755201
PDF version
Springer__Explainable Artificial Intelligence Based on Neuro-Fuzzy Modeling with.pdf
(10.99 MB, 需要: 30 个论坛币)
EPUB version
Springer__Explainable Artificial Intelligence Based on Neuro-Fuzzy Modeling with.epub
(40.92 MB, 需要: 30 个论坛币)
提升卡
置顶卡
沉默卡
变色卡
抢沙发
千斤顶
显身卡
京公网安备 11010802022788号
