Multidimensional Mining of Massive Text Data

Description

Unstructured text, as one of the most important data forms, plays a crucial role in data-driven decision making in domains ranging from social networking and information retrieval to scientific research and healthcare informatics. In many emerging applications, people's information need from text data is becoming multidimensional-they demand useful insights along multiple aspects from a text corpus. However, acquiring such multidimensional knowledge from massive text data remains a challenging task. This book presents data mining techniques that turn unstructured text data into multidimensional knowledge. We investigate two core questions. (1) How does one identify task-relevant text data with declarative queries in multiple dimensions? (2) How does one distill knowledge from text data in a multidimensional space? To address the above questions, we develop a text cube framework. First, we develop a cube construction module that organizes unstructured data into a cube structure, by discovering latent multidimensional and multi-granular structure from the unstructured text corpus and allocating documents into the structure. Second, we develop a cube exploitation module that models multiple dimensions in the cube space, thereby distilling from user-selected data multidimensional knowledge. Together, these two modules constitute an integrated pipeline: leveraging the cube structure, users can perform multidimensional, multigranular data selection with declarative queries; and with cube exploitation algorithms, users can extract multidimensional patterns from the selected data for decision making. The proposed framework has two distinctive advantages when turning text data into multidimensional knowledge: flexibility and label-efficiency. First, it enables acquiring multidimensional knowledge flexibly, as the cube structure allows users to easily identify task-relevant data along multiple dimensions at varied granularities and further distill multidimensional knowledge. Second, the algorithms for cube construction and exploitation require little supervision; this makes the framework appealing for many applications where labeled data are expensive to obtain. Overview......Page 16 Part I: Cube Construction......Page 18 Example Applications......Page 20 Technical Roadmap......Page 21 Task 1: Taxonomy Generation......Page 22 Task 3: Multidimensional Summarization......Page 23 Task 5: Abnormal Event Detection......Page 24 Organization......Page 25 Cube Construction Algorithms......Page 26 Overview......Page 28 Pattern-Based Extraction......Page 31 Clustering-Based Taxonomy Construction......Page 32 Adaptive Term Clustering......Page 33 Spherical Clustering for Topic Splitting......Page 34 Identifying Representative Terms......Page 35 Learning Local Term Embeddings......Page 37 Experimental Setup......Page 38 Qualitative Results......Page 40 Quantitative Analysis......Page 42 Summary......Page 45 Overview......Page 46 Related Work......Page 48 Framework Overview......Page 49 Hierarchical Tree Expansion......Page 50 Taxonomy Global Optimization......Page 55 Experimental Setup......Page 57 Qualitative Results......Page 58 Quantitative Results......Page 60 Summary......Page 63 Overview......Page 64 Latent Variable Models......Page 66 Preliminaries......Page 67 Modeling Class Distribution......Page 68 Generating Pseudo-Documents......Page 70 Neural Models with Self-Training......Page 71 Neural Model Self-Training......Page 72 Instantiating with CNNs and RNNs......Page 73 Experiments......Page 74 Baselines......Page 75 Experiment Settings......Page 76 Experiment Results......Page 77 Parameter Study......Page 79 Case Study......Page 82 Summary......Page 84 Overview......Page 86 Hierarchical Text Classification......Page 88 Pseudo-Document Generation......Page 89 Global Classifier Self-Training......Page 92 Algorithm Summary......Page 94 Experiment Settings......Page 95 Component-Wise Evaluation......Page 98 Summary......Page 101 Cube Exploitation Algorithms......Page 104 Introduction......Page 106 Preliminaries......Page 109 Text Cube Preliminaries......Page 110 Problem Definition......Page 111 Popularity and Integrity......Page 112 Neighborhood-Aware Distinctiveness......Page 113 Overview......Page 116 Hybrid Offline Materialization......Page 117 Optimized Online Processing......Page 121 Experimental Setup......Page 122 Effectiveness Evaluation......Page 123 Efficiency Evaluation......Page 127 Summary......Page 130 Overview......Page 132 Related Work......Page 134 Method Overview......Page 135 The Unsupervised Reconstruction Task......Page 137 The Optimization Procedure......Page 139 Life-Decaying Learning......Page 140 Constraint-Based Learning......Page 141 Experiments......Page 144 Experimental Setup......Page 145 Quantitative Comparison......Page 147 Case Studies......Page 149 Effects of Parameters......Page 152 Downstream Application......Page 154 Summary......Page 156 Overview......Page 158 Bursty Event Detection......Page 160 Preliminaries......Page 161 Method Overview......Page 162 Multimodal Embedding......Page 163 Candidate Generation......Page 165 A Bayesian Mixture Clustering Model......Page 166 Parameter Estimation......Page 167 Features Induced from Multimodal Embeddings......Page 168 Complexity Analysis......Page 169 Experimental Settings......Page 170 Qualitative Results......Page 172 Quantitative Results......Page 175 Scalability Study......Page 176 Summary......Page 177 Summary......Page 180 Future Work......Page 181 Bibliography......Page 184 Authors' Biographies......Page 198