
Advances in Knowledge Discovery and Data Mining
Description
Alles über E-Books | Antworten auf Fragen rund um E-Books, Kopierschutz und Dateiformate finden Sie in unserem Info- & Hilfebereich.
The three-volume set LNAI 11439, 11440, and 11441 constitutes the thoroughly refereed proceedings of the 23rd Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2019, held in Macau, China, in April 2019.
The 137 full papers presented were carefully reviewed and selected from 542 submissions. The papers present new ideas, original research results, and practical development experiences from all KDD related areas, including data mining, data warehousing, machine learning, artificial intelligence, databases, statistics, knowledge engineering, visualization, decision-making systems, and the emerging applications. They are organized in the following topical sections: classification and supervised learning; text and opinion mining; spatio-temporal and stream data mining; factor and tensor analysis; healthcare, bioinformatics and related topics; clustering and anomaly detection; deep learning models and applications; sequential pattern mining; weakly supervised learning; recommender system; social network and graph mining; data pre-processing and featureselection; representation learning and embedding; mining unstructured and semi-structured data; behavioral data mining; visual data mining; and knowledge graph and interpretable data mining.More details
Other editions
Additional editions

Persons
Content
- Intro
- PC Chairs' Preface
- General Chairs' Preface
- Organization
- Contents - Part II
- Deep Learning Models and Applications
- Semi-interactive Attention Network for Answer Understanding in Reverse-QA
- 1 Introduction
- 2 Related Work
- 2.1 QA
- 2.2 Text Classification
- 3 Methodology
- 3.1 Answer Understanding for T/F rQA
- 3.2 Answer Understanding for MC rQA
- 4 Experimental Data Construction
- 5 Experiment
- 5.1 Comparative Methods
- 5.2 Training Settings
- 5.3 Overall Competing Results
- 5.4 Discussion on the Key Modules in Our Models
- 6 Conclusion
- References
- Neural Network Based Popularity Prediction by Linking Online Content with Knowledge Bases
- 1 Introduction
- 2 Related Work
- 3 Problem Definition
- 4 The Proposed Model
- 4.1 A LSTM-Based Popularity Prediction Model
- 4.2 Enhancing the Prediction with KB Embeddings
- 4.3 Enhancing the Prediction with KB Neighbors
- 5 Experiments and Analysis
- 5.1 Experimental Setup
- 5.2 Results and Analysis
- 6 Conclusion
- References
- Passenger Demand Forecasting with Multi-Task Convolutional Recurrent Neural Networks
- 1 Introduction
- 2 Proposed Approach
- 2.1 Problem Formulation
- 2.2 Multi-Task CRNN (MT-CRNN) Framework
- 3 Experiments
- 3.1 Dataset
- 3.2 Experiment Settings
- 3.3 Experimental Results
- 4 Related Works
- 5 Conclusions
- References
- Accurate Identification of Electrical Equipment from Power Load Profiles
- 1 Introduction
- 2 Our Approach and Model Architecture
- 2.1 Generative Model
- 2.2 Discriminative Model
- 2.3 Loss Function
- 3 Experiments: Validate the Competition of Our Model
- 3.1 Baseline Methods
- 3.2 Datasets
- 3.3 Hyper Parameters
- 3.4 Experimental Results on UCR Datasets
- 4 Electrical Equipment Identification from Power Load Profiles
- 4.1 Dataset
- 4.2 Experimental Results
- 4.3 Training Process Analysis
- 5 Discussion and Conclusions
- References
- Similarity-Aware Deep Attentive Model for Clickbait Detection
- 1 Introduction
- 2 Related Work
- 2.1 Clickbait Detection
- 2.2 Deep Semantic Similarity Model
- 3 Methodology
- 3.1 Learn Latent Representations
- 3.2 Learn the Similarities
- 3.3 Learn for Prediction
- 4 Experiments
- 4.1 Dataset Description
- 4.2 Comparison Methods
- 4.3 Sensitivity Analysis
- 5 Conclusions
- References
- Topic Attentional Neural Network for Abstractive Document Summarization
- 1 Introduction
- 2 Related Work
- 3 Our Model
- 3.1 Overview
- 3.2 Paired Encoder
- 3.3 Paired-Attentional Decoder
- 3.4 Copying and Coverage
- 3.5 Topic Selection
- 4 Experiments
- 4.1 Dataset
- 4.2 Topic Information Acquisition
- 4.3 Implementation
- 4.4 Results and Discussion
- 5 Conclusion
- References
- Parameter Transfer Unit for Deep Neural Networks
- 1 Introduction
- 2 Related Works
- 3 Parameter Transfer Unit (PTU)
- 3.1 Three Transfer States
- 3.2 PTU for CNNs
- 3.3 PTU for RNNs
- 3.4 Scalability
- 4 Experimental Results
- 4.1 Experiments on CNNs
- 4.2 Experiments on RNNs
- 4.3 Convergence Performance
- 4.4 Integrate PTU with Feature-Based Transfer Learning Method
- 5 Conclusion
- References
- EFCNN: A Restricted Convolutional Neural Network for Expert Finding
- 1 Introduction
- 2 Problem Formulation
- 3 Our Model
- 3.1 Word Embedding and Similarity Matrix
- 3.2 EFCNN: Expert Finding with Restricted CNN
- 4 Experiments
- 4.1 Experimental Setup
- 4.2 Results and Discussion
- 5 Conclusions
- References
- CRESA: A Deep Learning Approach to Competing Risks, Recurrent Event Survival Analysis
- 1 Introduction
- 2 Cause-Specific Recurrent Event Survival Analysis
- 2.1 Survival Data
- 2.2 Model Description
- 2.3 Loss Function
- 3 Experiments
- 3.1 Dataset I: MIMIC III Clinical Dataset
- 3.2 Dataset II: Engine Failures Dataset
- 3.3 Dataset III: Synthetic Dataset
- 3.4 Training Details
- 3.5 Baselines
- 3.6 Performance Metrics and Results
- 4 Conclusions
- References
- Long-Term Traffic Time Prediction Using Deep Learning with Integration of Weather Effect
- 1 Introduction
- 2 Related Work
- 2.1 Real-Time and Short-Term Traffic Time Prediction
- 2.2 Long-Term Traffic Time Prediction
- 3 Proposed Framework
- 3.1 Dataset Pre-processing
- 3.2 Model Training
- 4 Experiment Evaluation
- 4.1 Data Description
- 4.2 Evaluation Metrics
- 4.3 Experimental Results
- 5 Conclusions
- References
- Arrhythmias Classification by Integrating Stacked Bidirectional LSTM and Two-Dimensional CNN
- Abstract
- 1 Introduction
- 2 Method
- 2.1 The Arrhythmias Classification Model Framework
- 2.2 The Wavelet Layer (WL)
- 2.3 The Stacked Bidirectional LSTM (SB-LSTM) Layer
- 2.4 The Two-Dimensional CNN (TD-CNN) Layer
- 2.5 The Fusion Layer (FL)
- 3 Experiment Evaluation
- 3.1 Dataset Description and Data Preprocessing
- 3.2 Experimental Setup
- 3.3 Evaluation Results
- 4 Conclusion and Future Work
- Acknowledgements
- References
- An Efficient and Resource-Aware Hashtag Recommendation Using Deep Neural Networks
- Abstract
- 1 Introduction
- 2 Preliminaries
- 2.1 Related Work
- 2.2 Convolutional Neural Network
- 2.3 Semantic Embedding Model
- 3 Model Architecture
- 4 Data
- 4.1 Image Dataset
- 4.2 Hashtag Dataset
- 5 Experiments and Results
- 5.1 Sampling of Image-Hashtag Pairs
- 5.2 Predicting and Recommending Hashtags
- 5.3 Verification and Inspection
- 6 Conclusion
- References
- Dynamic Student Classiffication on Memory Networks for Knowledge Tracing
- 1 Introduction
- 2 Knowledge Tracing
- 2.1 Bayesian Knowledge Tracing (BKT)
- 2.2 Deep Knowledge Tracing (DKT)
- 2.3 Dynamic Key-Value Memory Network (DKVMN)
- 2.4 Deep Knowledge Tracing with Dynamic Student Classification (DKT-DSC)
- 3 Dynamic Student Classification on Memory Networks (DSCMN)
- 3.1 Assessing Student's Mastery of Skill
- 4 Datasets
- 5 Experimental Study
- 6 Conclusion and Future Work
- References
- Targeted Knowledge Transfer for Learning Traffic Signal Plans
- 1 Introduction
- 2 Related Work
- 2.1 Approaches for Traffic Signal Control
- 2.2 Methods for Knowledge Transfer
- 3 Problem Definition
- 4 Method
- 4.1 Non-transfer Reinforcement Learning Solution
- 4.2 TTRL-B: Targeted Transfer Reinforcement Learning in a Batch Learning Framework
- 5 Experiments
- 5.1 Datasets
- 5.2 Compared Methods
- 5.3 Evaluation Metric
- 5.4 Overall Performance
- 5.5 Variants of Our Model
- 5.6 Parameter Sensitivity
- 5.7 Case Study of the Batch Learning Framework
- 6 Conclusion
- References
- Sequential Pattern Mining
- Efficiently Finding High Utility-Frequent Itemsets Using Cutoff and Suffix Utility
- 1 Introduction
- 2 Related Work
- 3 Proposed Model: High Utility-Frequent Itemset
- 4 EFIM and Its Limitations
- 5 Proposed Algorithm
- 5.1 Finding Secondary Items
- 5.2 Finding Candidate Items
- 5.3 Finding Primary Items
- 5.4 Recursive Mining of Primary Items
- 6 Experimental Results
- 7 Conclusions and Future Work
- References
- How Much Can A Retailer Sell? Sales Forecasting on Tmall
- 1 Introduction
- 2 Data Analysis and Problem Definition
- 2.1 Sales Data and Feature Description
- 2.2 Seasonality Analysis
- 2.3 Sale Amount Analysis
- 2.4 Problem Definition
- 3 Model Design and Implementation
- 3.1 Seasonality over Groups of Retailers
- 3.2 Tweedie Loss for Regression
- 4 Empirical Study
- 4.1 Dataset
- 4.2 Experimental Settings
- 4.3 Comparison Results
- 4.4 Effect of Tweedie Distribution Parameter ()
- 5 Related Works
- 5.1 Linear Model
- 5.2 Non-linear Model
- 6 Conclusions
- References
- Hierarchical LSTM: Modeling Temporal Dynamics and Taxonomy in Location-Based Mobile Check-Ins
- 1 Introduction
- 2 Problem Definition
- 2.1 Datasets: Foursquare and Jiepang
- 2.2 Observations
- 2.3 Problem Formulation
- 2.4 Data Preprocessing
- 3 Methodology
- 3.1 Hierarchical LSTM
- 4 Experimental Result
- 4.1 Experimental Setup
- 4.2 Models Compared and Previous Work
- 4.3 Result Summary
- 4.4 Taxonomy Embedding Analysis
- 5 Conclusion
- References
- Recovering DTW Distance Between Noise Superposed NHPP
- 1 Introduction
- 2 Related Work
- 3 Methods
- 3.1 DTW Distance on Sequences of Timestamps
- 3.2 Remove Noise Before DTW Calculation
- 3.3 Integrating Noise Removal Probability to DTW
- 4 Experiments
- 4.1 Synthetic Data
- 4.2 Classification on Real Data
- 4.3 Case Study for Customer Behaviour Segmentation
- 5 Conclusion
- References
- ATNet: Answering Cloze-Style Questions via Intra-attention and Inter-attention
- 1 Introduction
- 2 Related Work
- 2.1 LSTM with Attention
- 2.2 Pointer-Style Attention Sum
- 2.3 Self-attention
- 2.4 Multi-hop Architecture
- 3 ATNet
- 3.1 Contextual Encoding Representations
- 3.2 Intra-attention Aligner
- 3.3 Inter-attention Aligner
- 3.4 Answer Prediction Module
- 4 Experiments
- 4.1 Experimental Setups
- 4.2 Overall Results
- 5 Ablation Study
- 5.1 Effectiveness of Self-attention Module
- 5.2 Effectiveness of extAttention
- 6 Case Study
- 7 Conclusion
- References
- Parallel Mining of Top-k High Utility Itemsets in Spark In-Memory Computing Architecture
- Abstract
- 1 Introduction
- 2 Preliminary
- 3 Related Work
- 3.1 High Utility Itemset Mining
- 3.2 Top-K High Utility Itemset Mining
- 3.3 Parallel Mining of High Utility Patterns
- 4 The Proposed Method
- 4.1 Pre-evaluation in Parallel
- 4.2 Reorganize Transactions in Parallel
- 4.3 Mining Patterns in Parallel
- 5 Experimental Results
- 6 Conclusion
- Acknowledgement
- References
- Weakly Supervised Learning
- Robust Semi-supervised Multi-label Learning by Triple Low-Rank Regularization
- 1 Introduction
- 2 Related Work
- 3 Problem Formulation
- 3.1 Notations
- 3.2 The Regularization Framework
- 3.3 Optimization
- 4 Experiments
- 4.1 Datasets
- 4.2 Evaluation Criteria
- 4.3 Comparison with the State-of-the-art Algorithms
- 4.4 Multi-label Learning with Incomplete Labels
- 4.5 Parameters Sensitivity Analysis
- 5 Conclusion
- References
- Multi-class Semi-supervised Logistic I-RELIEF Feature Selection Based on Nearest Neighbor
- 1 Introduction
- 2 Proposed Method: MSLIR-NN
- 2.1 Margin Vectors
- 2.2 Optimization Problem
- 2.3 Algorithm and Complexity Analysis
- 2.4 Connections to Related Work
- 3 Experiments
- 3.1 Experiments on Multi-class Datasets
- 3.2 Experiments on Binary Datasets
- 3.3 Comparison of MSLIR-NN and MSLIR
- 4 Conclusions
- References
- Effort-Aware Tri-Training for Semi-supervised Just-in-Time Defect Prediction
- 1 Introduction
- 2 Related Work
- 2.1 Just-in-Time Defect Prediction
- 2.2 Semi-supervised Learning in Traditional Defect Prediction
- 3 Effort-Aware Tri-Training
- 3.1 Problem Formulation
- 3.2 Model Evaluation
- 3.3 Sample Selection
- 3.4 Result Prediction
- 4 Experiment Setup
- 4.1 Datasets
- 4.2 Baseline Models
- 4.3 Classifiers Selection
- 4.4 Evaluation Strategy
- 4.5 Performance Indicators
- 4.6 Research Questions
- 5 Experimental Results and Analysis
- 6 Conclusion
- References
- One Shot Learning with Margin
- 1 Introduction
- 2 One Shot Learning with Margin
- 2.1 One Shot Learning
- 2.2 One Shot Learning with Margin
- 3 Case Study
- 3.1 Prototypical Networks
- 3.2 Matching Networks
- 4 Experiments
- 4.1 Settings
- 4.2 Results on Omniglot
- 4.3 Results on miniImageNet
- 4.4 Parameter Study
- 5 Related Work
- 5.1 One Shot Learning
- 5.2 Metric Learning
- 6 Conclusion
- References
- DeepReview: Automatic Code Review Using Deep Multi-instance Learning
- 1 Introduction
- 2 The DeepReview Approach
- 2.1 The Framework of DeepReview
- 2.2 Data Processing
- 2.3 Instance Feature Generation Layer
- 2.4 Multi-instance Based Prediction Layer
- 3 Experiments
- 3.1 Experiment Settings
- 3.2 Experiment Results
- 4 Related Work
- 5 Conclusion
- References
- Multi-label Active Learning with Error Correcting Output Codes
- 1 Introduction
- 2 Related Work
- 3 Revisiting ECOC via Multi-label SVM Classification
- 4 The Algorithm
- 4.1 Max-Margin Uncertainty Sampling
- 4.2 Label Cardinality Inconsistency
- 4.3 Active Selection
- 5 Experiments
- 5.1 Experiments Setup
- 5.2 Comparison Results
- 5.3 Influence Analysis of Parameter
- 5.4 Compared with Traditional ECOC Method
- 6 Conclusion
- References
- Dynamically Weighted Multi-View Semi-Supervised Learning for CAPTCHA
- 1 Introduction
- 2 Related Work
- 3 Method
- 3.1 Filter Artificial Bee Colony Method
- 3.2 Dynamically Weighted Multi-View Semi-Supervised Learning
- 3.3 The Complete Algorithm
- 4 Experiments
- 4.1 Data Specification
- 4.2 Baselines
- 4.3 Result and Discussions on Feature Selection
- 4.4 Result and Discussions on DWMVSSL
- 5 Conclusion
- References
- Recommender System
- A Novel Top-N Recommendation Approach Based on Conditional Variational Auto-Encoder
- 1 Introduction
- 2 Preliminary
- 2.1 Variational Auto-encoder
- 2.2 Problem Description
- 3 Proposed Method
- 3.1 CVAE Model
- 3.2 The Split-Merge Framework for Multiple Conditions
- 4 Experiments
- 4.1 The Projection of Latent Feature
- 4.2 The Impact of Side Information
- 4.3 Performance Comparison
- 5 Related Works
- 6 Conclusion
- References
- Jaccard Coefficient-Based Bi-clustering and Fusion Recommender System for Solving Data Sparsity
- 1 Introduction
- 2 Jaccard Coefficient-Based Bi-clustering and Fusion
- 2.1 Motivation
- 2.2 JC-BiFu
- 3 Experiments
- 3.1 Datasets
- 3.2 Evaluation Metrices
- 3.3 Parameters Analysis
- 3.4 Performance
- 4 Conclusion
- References
- A Novel KNN Approach for Session-Based Recommendation
- 1 Introduction
- 2 Our Approach
- 2.1 Contextual KNN Approach
- 2.2 Candidate Selection
- 2.3 Diffusion-Based Similarity
- 3 Experiment
- 3.1 Experiment Setting
- 3.2 The Performance of Candidate Selection Strategy
- 3.3 The Study of and in DSM
- 3.4 Overall Performance
- 4 Conclusions
- References
- A Contextual Bandit Approach to Personalized Online Recommendation via Sparse Interactions
- 1 Introduction
- 2 Problem Formulation and Methodology
- 3 Our Approach
- 3.1 Estimating the User Attention
- 3.2 Learning the User Preference
- 3.3 Putting Everything Together
- 4 Regret Analysis
- 5 Experiments
- 6 Related Work
- 7 Conclusions
- References
- Heterogeneous Item Recommendation for the Air Travel Industry
- 1 Introduction
- 2 Related Work
- 3 Approach
- 3.1 The Air Travel HIN
- 3.2 The Joint Factorization Model
- 3.3 The Weighting Strategy
- 3.4 Model Optimization
- 4 Experiments
- 4.1 Dataset
- 4.2 Baselines
- 4.3 Experimental Settings
- 4.4 Experimental Results
- 4.5 Parameter Analysis
- 5 Conclusion
- References
- A Minimax Game for Generative and Discriminative Sample Models for Recommendation
- 1 Introduction
- 2 UGAN Formulation
- 2.1 A Minimax Sample Generation Framework
- 2.2 Loss Function
- 2.3 Extension to a Specific Case
- 3 Experiments and Analysis
- 3.1 Datasets and Evaluation Metrics
- 3.2 Experimental Results
- 4 Conclusion and Future Work
- References
- RNE: A Scalable Network Embedding for Billion-Scale Recommendation
- 1 Introduction
- 2 Related Work
- 2.1 Collaborative Filtering
- 2.2 Network Representation Learning
- 3 The Methodology
- 3.1 Network Embedding for Recommendation
- 3.2 Recommendation-Based Sub-sampling
- 3.3 Implementation
- 4 Experiments
- 4.1 Datasets
- 4.2 Online A/B Tests
- 4.3 Showcase
- 4.4 Offline Experiment
- 5 Conclusion
- References
- Social Network and Graph Mining
- Graph Compression with Stars
- 1 Introduction
- 2 The Star-Based Graph Compression
- 2.1 Star-Based Compressed Graphs
- 2.2 Star-Based Graph Compression Algorithm
- 3 Query Processing on Star-Based Compressed Graphs
- 3.1 Single-Source Shortest Path Queries
- 4 Experiments
- 4.1 Experimental Setting
- 4.2 Datasets
- 4.3 Performance of the Star-Based Graph Compression
- 4.4 Query Processing Performance on Star-Based Compressed Graphs
- 5 Related Work
- 6 Conclusions
- References
- Neighbor-Based Link Prediction with Edge Uncertainty
- 1 Introduction
- 2 Problem Definition
- 2.1 Uncertain Network
- 2.2 Link Prediction Problem Definition
- 3 Previous Work
- 4 Link Prediction for Uncertain Graphs
- 4.1 Time Complexity Analysis for the Calculation of Common Neighbors in Uncertain Networks
- 4.2 Time Complexity Analysis for the Calculation of Resource Allocation in Uncertain Networks
- 4.3 An Efficient Algorithm for the Calculation of Resource Allocation
- 5 Experiments
- 5.1 Datasets
- 5.2 Experiments
- 5.3 Results and Evaluation
- 6 Conclusion
- References
- Inferring Social Bridges that Diffuse Information Across Communities
- 1 Introduction
- 2 Related Work
- 3 Formalization
- 4 The Proposed Method
- 4.1 Utility of Bridge Links
- 4.2 Biased Features
- 4.3 The iBridge Framework
- 5 Experiment
- 5.1 Datasets and Settings
- 5.2 Comparative Methods
- 5.3 Results
- 5.4 Discussion
- 6 Conclusion
- References
- Learning Pretopological Spaces to Extract Ego-Centered Communities
- 1 Introduction
- 2 Related Works
- 3 Basics of Pretopology
- 4 Community Extraction Method
- 4.1 Optimization Functions Targeted at Pretopological Spaces Learning
- 4.2 From Network Descriptors to Predicates
- 4.3 Learning a Pretopological Space
- 4.4 Community Extraction from a Pretopological Space
- 5 Experiments
- 5.1 Datasets
- 5.2 Experimental Setup and Results
- 6 Conclusion
- References
- EigenPulse: Detecting Surges in Large Streaming Graphs with Row Augmentation
- 1 Introduction
- 2 Related Work
- 2.1 Anomaly Detection in Static Graphs
- 2.2 Anomaly Detection in Streaming Graphs
- 3 Proposed Method
- 3.1 Our Model
- 3.2 AugSVD Algorithm
- 3.3 EigenPulse Algorithm
- 4 Experiments
- 4.1 Experimental Settings
- 4.2 Q1.Efficiency
- 4.3 Q2.Accuracy
- 4.4 Q3.Scalability
- 5 Conclusion
- References
- TPLP: Two-Phase Selection Link Prediction for Vertex in Graph Streams
- 1 Introduction
- 2 Preliminaries
- 2.1 Link Prediction for Vertex in Graph Streams
- 2.2 Vertex-Biased Sampling
- 3 TPLP: A Two-Phase Selection Streaming Link Prediction Framework
- 3.1 Inverted Graph Sketch
- 3.2 Two-Phase Selection Algorithm
- 3.3 Estimation of Common Neighbor
- 4 Experiments
- 4.1 Datasets
- 4.2 Performance of Two-Phase Selection
- 4.3 Performance of Common Neighbor Estimation
- 4.4 Performance of Link Prediction Accuracy
- 4.5 Error Incurred by Sampling
- 5 Conclusions
- References
- Robust Temporal Graph Clustering for Group Record Linkage
- 1 Introduction
- 2 Related Work
- 3 Overview of Temporal Graph Clustering
- 4 Temporal Connected Component Clustering
- 5 Iterative Cluster Merging
- 6 Experimental Evaluation
- 7 Conclusions and Future Work
- References
- Data Pre-processing and Feature Selection
- Learning Diversified Features for Object Detection via Multi-region Occlusion Example Generating
- 1 Introduction
- 2 Related Work
- 3 Our Method
- 3.1 Multi-region Occlusion Example Generating
- 3.2 Implementation Details
- 4 Experiments
- 4.1 Experimental Settings
- 4.2 Ablative Analysis
- 4.3 Comparisons with A-Fast-RCNN and OHEM
- 4.4 PASCAL VOC2012 and MS COCO Results
- 4.5 Visualization
- 5 Conclusions
- References
- HATDC: A Holistic Approach for Time Series Data Repairing
- 1 Introduction
- 2 Related Work
- 3 Anomaly Detection
- 4 Dirty Data Repairing
- 5 Clustering-Based Optimization
- 6 Experiments
- 6.1 Accuracy of Anomaly Detection
- 6.2 Comparison with Existing Approaches
- 6.3 Evaluation on Various Window Size
- 6.4 Evaluation on Various Cluster Number
- 7 Conclusion
- References
- Double Weighted Low-Rank Representation and Its Efficient Implementation
- 1 Introduction
- 2 Proposed Double Weighted Model
- 2.1 Weighted Feature Learning for Error Penalizing
- 2.2 Weighted Rational Function for Rank Approximation
- 3 Optimization Algorithm
- 3.1 Reweighted Framework
- 3.2 Accelerated Proximal Gradient Algorithm
- 3.3 Automatic Singular Value Thresholding
- 3.4 Efficient SVD
- 3.5 Complexity and Convergence
- 4 Experimental Results
- 4.1 Clustering Performance
- 4.2 Execution Time
- 5 Conclusions
- References
- Exploring Dual-Triangular Structure for Efficient R-Initiated Tall-Skinny QR on GPGPU
- 1 Introduction
- 2 Related Work
- 2.1 Householder QR
- 2.2 Givens QR
- 2.3 Cholesky QR
- 2.4 TSQR
- 3 R-Initiated Tall-Skinny QR on GPGPU
- 3.1 R-Initiated Method to Meet the Memory Limitation of GPGPU
- 3.2 Dual-Triangular Structure to Accelerate the Process
- 4 Experimental Results
- 4.1 Results for Different Algorithms
- 4.2 Sensitivity of Number of Column N
- 4.3 Sensitivity of Row-Column Ratio M
- 4.4 The Bottleneck: Computing Normal QR
- 4.5 PCA with Tall-Skinny QR
- 5 Conclusion
- References
- Efficient Autotuning of Hyperparameters in Approximate Nearest Neighbor Search
- 1 Introduction
- 2 Approximate Nearest Neighbor Search
- 3 Randomized Space-Partitioning Trees
- 3.1 Index Construction
- 3.2 ANN Search Using Multiple Trees
- 3.3 Comparison of Randomization and Search Methods
- 4 An Autotuning Algorithm
- 4.1 Estimating Recall and Candidate Set Size
- 4.2 Estimating the Query Time
- 4.3 Using the Autotuning Index
- 5 Experimental Results
- References
- An Accelerator of Feature Selection Applying a General Fuzzy Rough Model
- Abstract
- 1 Introduction
- 2 Preliminaries
- 2.1 Fuzzy Rough Sets
- 3 Fuzzy Rough Based Feature Selection Accelerator
- 3.1 Some Theorems
- 3.2 Fuzzy Rough Based Feature Selection Accelerator
- 4 Experimental Analysis
- 4.1 Experimental Setup
- 4.2 Compare DAR and PAR
- 4.3 Compare PAR and PARA
- 4.4 The Classification Performance Comparison of Three Algorithms
- 5 Conclusions
- Acknowledgements
- References
- Text Feature Extraction and Selection Based on Attention Mechanism
- 1 Introduction
- 2 Our Proposed Model
- 2.1 Multiple Attention and Pooling Strategies
- 2.2 Encoder Layers
- 2.3 Second Attention Layer
- 2.4 Compare, Aggregate, Softmax Layers
- 3 Experiments
- 3.1 Datasets and Experimental Protocol
- 3.2 Experimental Results
- 4 Discussion and Analysis
- 4.1 Question Type Analysis
- 4.2 In-Depth Analysis
- 4.3 Ablation Analysis
- 5 Related Work
- 6 Conclusion
- References
- Author Index
System requirements
File format: PDF
Copy protection: Watermark-DRM (Digital Rights Management)
System requirements:
- Computer (Windows; MacOS X; Linux): Use the free software Adobe Reader, Adobe Digital Editions, or any other PDF viewer of your choice (see eBook Help).
- Tablet/Smartphone (Android; iOS): Install the free app Adobe Digital Editions or another reading app for eBooks, e.g., PocketBook (see eBook Help).
- E-reader: Bookeen, Kobo, Pocketbook, Sony, Tolino and many more (only limited: Kindle).
The file format PDF always displays a book page identically on any hardware. This makes PDF suitable for complex layouts such as those used in textbooks and reference books (images, tables, columns, footnotes). Unfortunately, on the small screens of e-readers or smartphones, PDFs are rather annoying, requiring too much scrolling.
This eBook uses Watermark-DRM, a „soft” copy protection. This means that there are no technical restrictions to prevent illegal distribution. However, there is a personalised watermark embedded in the eBook that can be used to identify the purchaser of the eBook in the event of misuse and to provide evidence for legal purposes.
For more information, see our eBook Help page.