Intelligent Computing Theories and Application: 17th International Conference, ICIC 2021, Shenzhen, China, August 12–15, 2021, Proceedings, Part III (Lecture Notes in Computer Science)

Description

This two-volume set of LNCS 12836 and LNCS 12837 constitutes - in conjunction with the volume LNAI 12838 - the refereed proceedings of the 17th International Conference on Intelligent Computing, ICIC 2021, held in Shenzhen, China in August 2021. The 192 full papers of the three proceedings volumes were carefully reviewed and selected from 458 submissions.  The ICIC theme unifies the picture of contemporary intelligent computing techniques as an integral concept that highlights the trends in advanced computational intelligence and bridges theoretical research with applications. The theme for this conference is “Advanced Intelligent Computing Methodologies and Applications.” The papers are organized in the following subsections: Artificial Intelligence in Real World Applications, Biomedical Informatics Theory and Methods, Complex Diseases Informatics, Gene Regulation Modeling and Analysis, Intelligent Computing in Computational Biology, and Protein Structure and Function Prediction. Preface Organization Contents – Part III Artificial Intelligence in Real World Applications Task-Oriented Snapshot Network Construction of Stock Market 1 Introduction 2 Methods of Snapshot Network Construction 2.1 Formulation of Task-Oriented Problem 2.2 Snapshot Network Construction with SHAP 3 Experiments and Results 4 Discussion and Conclusions References Analysis of Elimination Algorithm Based on Curve Self-intersection 1 Introduction 2 Parametric Representation of the Offset Curve 2.1 Equidistant Offset Curves 2.2 Variable Offset Curves 3 The Self-intersection Detection of Offset Curves 3.1 Intersecting Problem of Offset Curves 3.2 The solution of the Self-intersection Problem of Curves 4 Examples 4.1 Example Analysis of Power Base Representation 4.2 Example Analysis of Bezier Base Representation 5 Summary and Prospect References Towards AI-Based Reaction and Mitigation for e-Commerce - the ENSURESEC Engine 1 Context and Motivation 2 Vision of the Proposed Tool 3 Competitive Solutions 4 Technical Implementation 4.1 Input Data and Machine Learning Based Detection 4.2 Response and Mitigation Leveraging the Accumulated Knowledge 4.3 Maintaining Business Processes Intact 5 Conclusions and Future Work References Arabic Light Stemmer Based on ISRI Stemmer 1 Introduction 2 Literature Review 3 Proposed Methods 3.1 Preprocessing 4 Experimental Results 4.1 Dataset Description 4.2 Experiments and Result 4.3 Comparative Analysis 4.4 Comparison Between Proposed and Existing Light Stemming Methods 5 Discussion 6 Conclusion References Biomedical Informatics Theory and Methods Predicting miRNA-Disease Associations via a New MeSH Headings Representation of Diseases and eXtreme Gradient Boosting 1 Introduction 2 Methods and Materials 2.1 Human miRNA-Disease Associations 2.2 MiRNA Functional Similarity 2.3 Gaussian Interaction Profile (GIP) Kernel Similarity for miRNA and Disease 2.4 Disease Semantic Similarity 2.5 MeSHHeading2vec Method 2.6 Multi-source Feature Fusion 3 Experimental Results 4 Conclusions References Social Media Adverse Drug Reaction Detection Based on Bi-LSTM with Multi-head Attention Mechanism 1 Introduction 2 Method 2.1 Bi-LSTM 2.2 Multi-head Attention Mechanisms (MDAM) 3 Bi-LSTM with MDAM for Social Media ADR Detection 4 Experiments and Results 5 Conclusions References HOMC: A Hierarchical Clustering Algorithm Based on Optimal Low Rank Matrix Completion for Single Cell Analysis 1 Introduction 2 Method 2.1 Low Rank Matrix Completion 2.2 Hierarchical Clustering 3 Results 3.1 Data Materials 3.2 Performance Evaluation 4 Conclusion References mzMD: A New Storage and Retrieval System for Mass Spectrometry Data 1 Introduction 2 Methods 2.1 Mass Spectrometry Data Storage System 2.2 Query a Data Window for a Summary 3 Results 3.1 Query for Data Window Summaries 4 Conclusions References Drug-Target Interaction Prediction via Multiple Output Graph Convolutional Networks 1 Introduction 2 Proposed Method 2.1 Problem Description and Notation Definition 2.2 Overview of Our Method 2.3 Feature Extraction 2.4 Low-Level Features Extraction and Feature Concatenation 2.5 Graph Calculation 2.6 The Proposed MOGCN Model 2.7 Architectural Parameter 3 Experiments 3.1 Dataset 3.2 Compared Methods 3.3 Experimental Setting 3.4 The CVD Experiments 3.5 The CVT Experiments 3.6 The CVP Experiments 4 Conclusion References Inversion of k-Nearest Neighbours Algorithm for Extracting SNPs Discriminating Human Populations 1 Introduction 2 Method 2.1 The Convergence of Algorithm 3 Results 4 Conclusion References ComPAT: A Comprehensive Pathway Analysis Tools 1 Introduction 2 Materials and Methods 3 Results 4 Discussion and Conclusion References Incorporating Knowledge Base for Deep Classification of Fetal Heart Rate 1 Introduction 2 Method 2.1 Knowledge Base 2.2 Classification with Knowledge Base 3 Experiments Setup and Results 3.1 Dataset and Preprocessing 3.2 Implementation Details 3.3 Evaluation Metrics 3.4 Comparison Methods 3.5 Results Analysis 4 Conclusion References Review of Methods for Data Collection Experiments with People with Dementia and the Impact of COVID-19 1 Introduction 2 Methodology 3 Discussion 3.1 Recruitment 3.2 Consent and Assent Acquisition 3.3 Physiological Data Collection 3.4 Observational Data Collection 3.5 Data Transfer and Storage 4 Conclusion References KGRN: Knowledge Graph Relational Path Network for Target Prediction of TCM Prescriptions 1 Introduction 2 Related Works 3 Methods 3.1 Embedding Layer 3.2 Relational Path-Aware Aggregation Layer 3.3 Prediction Layer 3.4 Optimization 4 Experiments 4.1 Dataset 4.2 Experimental Settings 4.3 Performance Comparison 4.4 Hyper-parameters Study 5 Conclusions References Challenges in Data Capturing and Collection for Physiological Detection of Dementia-Related Difficulties and Proposed Solutions 1 Introduction 2 Background 3 Dataset Search Methodology 4 Data Search Results 5 Discussion of Solutions and Proposed Methodology 5.1 Data Collection Experiment 5.2 Anonymization and Pseudonymization 5.3 Synthetic Data 5.4 Model Training and Data Sharing 6 Conclusion References Exploring Multi-scale Temporal and Spectral CSP Feature for Multi-class Motion Imagination Task Classification 1 Introduction 2 Description of Dataset and CSP Feature Extraction 3 Multiscale Features Extracted by CSP 3.1 Multiscale Spectral CSP Features 3.2 Multiscale Temporal CSP Features 3.3 Multiscale Temporal and Spectral CSP Features 4 Results and Discussions 5 Conclusion References Gingivitis Detection by Wavelet Energy Entropy and Linear Regression Classifier 1 Introduction 2 Dataset 3 Methodology 3.1 Wavelet Energy Entropy 3.2 Linear Regression Classifier 3.3 10-fold Cross Validation 4 Experiment Results and Discussions 4.1 Wavelet Result 4.2 Statistical Analysis 4.3 Comparison to State-of-the-art Approaches 5 Conclusions References Decomposition-and-Fusion Network for HE-Stained Pathological Image Classification 1 Introduction 2 Method 2.1 Nuclei Segmentation 2.2 DFNet Architecture 3 Experiments 3.1 Datasets and Preprocessing 3.2 Classification Results 4 Conclusions References Complex Diseases Informatics A Novel Approach for Predicting Microbe-Disease Associations by Structural Perturbation Method 1 Introduction 2 Material and Methods 2.1 Datasets 2.2 Methods Overview 2.3 Disease Similarity Measure 2.4 Microbe Similarity Measure 2.5 Construction of Bi-layer Network 2.6 Structural Perturbation Model 3 Results 3.1 Model Design 3.2 Performance Evaluation 3.3 Compared with Other the-State-of-Art Methods 3.4 Case Studies 4 Conclusion References A Reinforcement Learning-Based Model for Human MicroRNA-Disease Association Prediction 1 Introduction 2 Materials and Methods 2.1 Human miRNA-Disease Associations 2.2 miRNA Functional Similarity 2.3 Disease Semantic Similarity 2.4 Method Models 3 Results 3.1 Evaluation Measurements 3.2 Comparison with Other Methods 4 Conclusion and Discussion References Delineating QSAR Descriptors to Explore the Inherent Properties of Naturally Occurring Polyphenols, Responsible for Alpha-Synuclein Amyloid Disaggregation Scheming Towards Effective Therapeutics Against Parkinson’s Disorder 1 Introduction 2 Methodology 2.1 Dataset 2.2 Polyphenols’ Structural Information 2.3 Computing Descriptors/Features of Polyphenols 2.4 Feature Selection 3 Results and Discussion 4 Conclusion References Study on the Mechanism of Cistanche in the Treatment of Colorectal Cancer Based on Network Pharmacology 1 Introduction 2 Materials and Methods 2.1 The Effective Ingredients and Targets of Cistanche 2.2 Drug-Disease Target Network Construction 2.3 Construction of Target Interaction Network and Screening of Core Genes 2.4 GO and KEGG Signal Pathway Enrichment Analysis 2.5 The Simulation of Drug Active Ingredient and Target Protein via Molecular Docking 3 Results 3.1 Screening of Active Ingredients and Prediction of the Target of Cistanche 3.2 Potential Targets of Cistanche in the Treatment of Colorectal Cancer 3.3 Construction of Active Pharmaceutical Ingredient-Disease Target Network 3.4 Construction and Analysis of Target Protein PPI Network 3.5 GO and KEGG Signal Pathway Enrichment Analysis 3.6 Molecular Docking 4 Discussion References A Novel Hybrid Machine Learning Approach Using Deep Learning for the Prediction of Alzheimer Disease Using Genome Data 1 Introduction 2 Methodology 2.1 Dataset 2.2 Quality Control 2.3 Association Test - Logistic Regression 2.4 Feature Selection 2.5 Classification 2.6 The Proposed Model 3 Results and Discussion 4 Conclusion and Future Work References Prediction of Heart Disease Probability Based on Various Body Function 1 Introduction 2 Methods and Materials 2.1 Data 3 Results and Discussion 4 Conclusion References Classification of Pulmonary Diseases from X-ray Images Using a Convolutional Neural Network 1 Introduction 2 State of the Art 3 Proposed Method 3.1 Convolutional Neural Network (CNN) 3.2 Data Set 4 Experimental Results 4.1 Network Training 4.2 Test 5 Conclusions References Predicting LncRNA-Disease Associations Based on Tensor Decomposition Method 1 Introduction 2 Method and Materials 2.1 Datasets 2.2 Tensor Integration 2.3 Construct Disease Matrix and LncRNA Matrix 2.4 Reconstruct Association Tensor 2.5 Optimization 3 Result 3.1 Experimental Setting and Evaluation Metrics 3.2 Parameter Analysis 3.3 Comparison with Other Existing Methods 4 Conclusions References AI in Skin Cancer Detection 1 Introduction 2 Skin Cancer 3 Skin Lesions Distribution: Age and Gender 4 Skin Lesions Sites 5 Applied Deep Learning for Detecting the Skin Cancer: Related Research 6 Conclusion References miRNA-Disease Associations Prediction Based on Neural Tensor Decomposition 1 Introduction 2 Method 2.1 MiRNA-Gene-Disease Association Tensor 2.2 Task Description 2.3 NTDMDA Method 3 Experiments 3.1 Data Collection 3.2 Performance Evaluation 3.3 Case Study 3.4 The Impact of Parameters 4 Conclusion References Gene Regulation Modeling and Analysis SHDC: A Method of Similarity Measurement Using Heat Kernel Based on Denoising for Clustering scRNA-seq Data 1 Introduction 2 Method 2.1 Framework Overview 2.2 Denoising Using DCA 2.3 Using Single Kernel Instead of Multi-Kernel 2.4 Selecting Heat Kernel Instead of Gaussian Kernel 3 Results 3.1 scRNA-seq Datasets 3.2 Clustering Accuracy 3.3 Parameter Sensitivity Analysis 3.4 Evaluation of Denoising 3.5 The Scalability of SHDC 3.6 Cell Visualization 3.7 Contributions of Denoising and Similarity Measure to SHDC 3.8 Implementation 4 Conclusion References Research on RNA Secondary Structure Prediction Based on MLP 1 Introduction 2 RNA Pseudoknot Structure and Label Introduction 2.1 RNA Pseudoknot Structure 2.2 RNA Pseudoknot Label 3 MLP Neural Network Prediction Model 4 Experiments and Analysis 4.1 Data Set 4.2 Evaluation Indicators 4.3 Analysis of Results 5 Conclusion References Inference of Gene Regulatory Network from Time Series Expression Data by Combining Local Geometric Similarity and Multivariate Regression 1 Introduction 2 Methods 2.1 Local Geometric Similarity 2.2 Multivariate Regression 2.3 Combining Local and Global Contributions 3 Results 3.1 In DREAM In-Silico Datasets 3.2 In HCC Datasets 4 Conclusion References Deep Convolution Recurrent Neural Network for Predicting RNA-Protein Binding Preference in mRNA UTR Region 1 Introduction 2 Methods 2.1 Encoding Layer 2.2 Convolution Layer 2.3 LSTM 2.4 Output Layer 3 Experiments 3.1 Datasets and Metrics 3.2 Results and Analysis 3.3 The Effects of Model Structure Changes on Model Performance 4 Conclusions References Joint Association Analysis Method to Predict Genes Related to Liver Cancer 1 Introduction 2 Methods and Materials 2.1 GO Analysis 2.2 KEGG Analysis 2.3 PPI Analysis 2.4 DO Analysis 2.5 Comparison with Other Methods 3 Conclusion References A Hybrid Deep Neural Network for the Prediction of In-Vivo Protein-DNA Binding by Combining Multiple-Instance Learning 1 Introduction 2 Materials and Methods 2.1 Benchmark Dataset 2.2 Methods 2.3 Evaluation Metrics 2.4 Hyper-parameter Settings 3 Result and Discussion 3.1 Competing Methods 3.2 Comparison K-mer 4 Conclusion References Using Deep Learning to Predict Transcription Factor Binding Sites Combining Raw DNA Sequence, Evolutionary Information and Epigenomic Data 1 Introduction 2 Materials and Methods 2.1 Dataset and Preprocessing 2.2 Network Architecture 2.3 Evaluation Metric 2.4 Experiment Setting 3 Results and Analysis 3.1 Results Display 3.2 Effect of Conservation Scores (Convs), MeDIP-seq (MDS), Histone Modifications (HMS) 4 Conclusion and Future Work References An Abnormal Gene Detection Method Based on Selene 1 Introduction 2 Related Work 2.1 Introduction to the Selene Framework 2.2 Acquisition and Processing of Data Sets 2.3 Introduction to Models 3 Experimental Details 3.1 Analysis of Results 4 Conclusion References A Method for Constructing an Integrative Network of Competing Endogenous RNAs 1 Introduction 2 Materials and Methods 2.1 Data Collection and Filtering 2.2 Deriving ceRNA-miRNA-ceRNA Interactions 2.3 Constructing an Integrative CeRNA Network 2.4 Finding Potential Prognostic Triplets 2.5 Constructing a Subnetwork of Potential Prognostic Triplets 3 Results and Discussion 3.1 Integrative ceRNA Network 3.2 Potential Prognostic Triplets 3.3 Subnetwork of Potential Prognostic Triplets 4 Conclusion References Intelligent Computing in Computational Biology Detection of Drug-Drug Interactions Through Knowledge Graph Integrating Multi-attention with Capsule Network 1 Introduction 2 Related Work 3 Materials and Methods 3.1 Dataset 3.2 Multi-attention Learning Strategy 3.3 Capsule Network 3.4 Knowledge Graph Representation Learning 3.5 Performance Evaluation Indicators 4 Experiments 4.1 Baseline 4.2 Prediction Performance of Proposed Method 4.3 Comparison of Several Types of Representation Learning Method 5 Conclusion References SCEC: A Novel Single-Cell Classification Method Based on Cell-Pair Ensemble Learning 1 Introduction 2 Materials and Method 2.1 Datasets 2.2 Dimension Reduction Methods 2.3 Unsupervised Base Classifiers 2.4 Ensemble Learning 2.5 Evaluation Metrics 3 Results and Discussion 3.1 Performance on PBMCs Dataset 3.2 Performance on Chu Cell Type Dataset 3.3 Performance on Klein Cell Type Dataset 3.4 Performance on Zeisel Cell Type Dataset 3.5 Visualization by Consensus Incidence Matrix 3.6 Discussion References ICNNMDA: An Improved Convolutional Neural Network for Predicting MiRNA-Disease Associations 1 Introduction 2 Feature Cell 2.1 Human MiRNA-Disease Associations 2.2 Disease Similarity 2.3 miRNA Similarity 3 ICNNMDA 3.1 Overview for ICNNMDA 3.2 Convolutional Layer 3.3 Pooling Layer 3.4 Fully-Connected Layer 3.5 Model Training 4 Experiments and Results 4.1 Parameter Setting 4.2 Performance Evaluation Based on 5CV 4.3 A Case Study on Prostate Neoplasm 5 Discussion and Conclusion References DNA-GCN: Graph Convolutional Networks for Predicting DNA-Protein Binding 1 Introduction 2 Related Work 2.1 Deep Learning for Motif Inference 2.2 Graph Convolutional Networks 2.3 Heterogeneous Graph 3 DNA-GCN 3.1 Sequence k-mer Graph 3.2 DNA-GCN 3.3 Implementation of DNA-GCN 4 Result 4.1 Datasets 4.2 Baselines 4.3 Our Model Outperforms on Many Datasets 5 Conclusion References Weighted Nonnegative Matrix Factorization Based on Multi-source Fusion Information for Predicting CircRNA-Disease Associations 1 Introduction 2 Materials and Methods 2.1 Dataset 2.2 Method Overview 2.3 Similarity Measures 2.4 Weighted K Nearest Neighbor Profiles for CircRNAs and Diseases 2.5 WNMFCDA 3 Results 4 Conclusions 5 Competing Interests References ScSSC: Semi-supervised Single Cell Clustering Based on 2D Embedding 1 Introduction 2 Dataset 3 Methods 3.1 Data Pre-processing 3.2 Image Synthesis 3.3 Pre-training with Autoencoder 3.4 Build a Network and Use Community Discovery Algorithms to Classify Data 3.5 Semi-supervised Neural Network Model 3.6 Evaluation Index 4 Results 4.1 Similarity Heat Map 4.2 Results of Clustering Performance Indicators (ARI and NMI) 4.3 Visualization of Results 4.4 Parameter Sensitivity 5 Conclusion References SNEMO: Spectral Clustering Based on the Neighborhood for Multi-omics Data 1 Introduction 2 Materials and Methods 2.1 Similar Matrix 2.2 Similarity Matrix Integration 2.3 K Value Selection and Clustering 2.4 Datasets 3 Results 3.1 The Results of Eigengap Heuristic 3.2 The Result of Custom K Value 3.3 A Case Study: Subtype Analysis in GBM 4 Discussion References Covid-19 Detection by Wavelet Entropy and Jaya 1 Introduction 2 Dataset 3 Methodology 3.1 Wavelet Entropy 3.2 Feedforward Neural Network 3.3 Jaya Algorithm 3.4 K-fold Cross Validation 4 Experiment Results and Discussions 4.1 WE Results 4.2 Statistical Results 4.3 Comparison to State-of-the-Art Approaches 5 Conclusions References An Ensemble Learning Algorithm for Predicting HIV-1 Protease Cleavage Sites 1 Introduction 2 Materials and Methods 2.1 Feature Vector Construction 2.2 Asymmetric Bagging SVM Classifier 3 Experiment Results 3.1 Experimental Datasets 3.2 Evaluation Metrics 3.3 10-Fold Cross Validation 3.4 Analysis of Feature Significance and Contribution 4 Conclusion References RWRNCP: Random Walking with Restart Based Network Consistency Projection for Predicting miRNA-Disease Association 1 Introduction 2 Materials and Methods 2.1 Human miRNA-Disease Association 2.2 miRNA Function Similarity 2.3 Disease Semantic Similarity 2.4 Topological Similarity of Disease and miRNA 2.5 Gaussian Interaction Profile Kernel Similarity for Diseases and miRNA 2.6 Integrated Similarity for Diseases and miRNAs 2.7 Network Consistency Projection of miRNA-Disease Association 3 Experiments and Results 3.1 Performance Evaluation 3.2 A Case Study 4 Discussion References MELPMDA: A New Method Based on Matrix Enhancement and Label Propagation for Predicting miRNA-Disease Association 1 Introduction 2 Materials 2.1 Human miRNA-Disease Associations 2.2 Disease Semantic Similarity 2.3 miRNA Functional Similarity 3 MELPMDA 3.1 Enhance the Similarity Matrix Through Similarity Reward Matrix (ESSRM) 3.2 Enhance the Association Matrix Through Self-adjusting Nearest Neighbor Method (EASNN) 3.3 Label Propagation (LP) 4 Experiments 4.1 Performance Evaluation 4.2 Case Study 4.3 Ablation Study 5 Conclusions References Prognostic Prediction for Non-small-Cell Lung Cancer Based on Deep Neural Network and Multimodal Data 1 Introduction 2 Materials and Methods 2.1 Dataset and Data Processing 2.2 Feature Transformation 2.3 Fusion Deep Neural Network 3 Experiments and Results 3.1 Performance of Feature Transformation Models 3.2 Performance of the Fusion Deep Neural Network 4 Discussion and Conclusion References Drug-Target Interactions Prediction with Feature Extraction Strategy Based on Graph Neural Network 1 Introduction 2 Method 2.1 Graph-Based Feature Extraction 2.2 GROAN-Based Feature Extraction Strategy 3 Experiments 3.1 Dataset 3.2 Experimental Results 4 Conclusion References CNNEMS: Using Convolutional Neural Networks to Predict Drug-Target Interactions by Combining Protein Evolution and Molecular Structures Information 1 Introduction 2 Materials and Methods 2.1 Benchmark Datasets 2.2 Drug Represented by Molecular Fingerprint Descriptor 2.3 Representing Target Protein with PSSM 2.4 Feature Extraction Using CNN 2.5 Classification by ELM 3 Results 3.1 Evaluation Criteria 3.2 Evaluate Prediction Performance 3.3 Comparison with Other Outstanding Methods 4 Conclusion References A Multi-graph Deep Learning Model for Predicting Drug-Disease Associations 1 Introduction 2 Materials and Methods 2.1 Experimental Dataset 2.2 The Input Feature Extraction of Graph Attention Network 2.3 Graph Attention Network 2.4 Graph Embedding Algorithm 3 Results and Discussion 3.1 The Evaluation Performance of the Proposed Model 3.2 The Proposed Model Compares with Other Methods 4 Conclusions References Predicting Drug-Disease Associations Based on Network Consistency Projection 1 Introduction 2 Materials and Methods 2.1 Dataset 2.2 Disease Similarity Based on Drug Characteristics 2.3 Gaussian Similarity of Drugs and Diseases 2.4 Integrated Similarity for Drugs and Diseases 2.5 Network Consistency Projection Method 3 Experiments 3.1 Evaluation Metrics 3.2 Parameter Analysis 3.3 Comparison with Other Methods 3.4 Comparison of Different Similarity Network Fusion Methods 3.5 Case Studies 4 Conclusion References An Efficient Computational Method to Predict Drug-Target Interactions Utilizing Matrix Completion and Linear Optimization Method 1 Introduction 2 Materials and Methods 2.1 Datasets 2.2 Similarity Measure of Drug and Target 2.3 Enhancing the Information of Similar Network 2.4 Construction of the Heterogeneous Network 2.5 Linear Optimization Model for Drug-Target Interactions Prediction 3 Result 3.1 Experimental Setting and Evaluation Metrices 3.2 Overall Performance 3.3 Comparison with Other Existing Methods 4 Conclusions References Protein Structure and Function Prediction Protein-Protein Interaction Prediction by Integrating Sequence Information and Heterogeneous Network Representation 1 Introduction 2 Materials and Methods 2.1 Dataset Collecting 2.2 Local Feature Extraction from Protein Sequence by K-mer 2.3 Global Feature Extraction from Network by LINE 2.4 Protein Representation 2.5 Performance Evaluation Indicators 3 Experiments 3.1 Prediction Performance of Proposed Method with Three Types of Representations 3.2 Comparison of Different Machine Learning Classifiers 4 Conclusion References DNA-Binding Protein Prediction Based on Deep Learning Feature Fusion 1 Introduction 2 Data Set and Data Representation 2.1 Data Set 2.2 Data Representation 3 Methods 3.1 Model Description 3.2 Model Parameters 4 Results and Discussion 5 Conclusion References Membrane Protein Identification via Multiple Kernel Fuzzy SVM 1 Introduction 2 Materials and Methods 2.1 Data Set 2.2 Extracting Evolutionary Conservatism Information 2.3 Support Vector Machine 2.4 Fuzzy Support Vector Machine 2.5 Multiple Kernel Support Vector Machine Classifier 3 Results 3.1 Evaluation Measurements 3.2 Prediction on Dataset 1 3.3 Prediction on Dataset 2 4 Conclusion and Discussion References Golgi Protein Prediction with Deep Forest 1 Introduction 2 Methods and Materials 2.1 Data 2.2 Construction of Classification 2.3 Evaluation Methods 3 Results and Discussion 3.1 Results 3.2 Discussion 4 Conclusion References Prediction of Protein-Protein Interaction Based on Deep Learning Feature Representation and Random Forest 1 Introduction 2 Materials and Methods 2.1 Datasets 2.2 Feature Extraction 3 Building a Random Forest Classifier 4 Results and Discussion 4.1 Evaluation of the Method 4.2 BiLSTM-RF Shows Better Performance 4.3 The Feature Information Extracted by Deep Learning is More Comprehensive 5 Conclusion References Author Index