Multimodal Sentiment Analysis (eBook)

Preface 

1 Introduction and MotivationResearch Challenges in Text-Based Sentiment AnalysisResearch Challenges in Multimodal Sentiment AnalysisOverview of the Proposed FrameworkContributions of this BookBook Organisation

 2 BackgroundAffective ComputingSentiment AnalysisPattern Recognition  Feature SelectionModel Evaluation Techniques  Model Validation TechniquesClassification TechniquesFeature-Based Text RepresentationConclusion

 3 Literature Survey and DatasetsIntroductionAvailable DatasetsVisual, Audio Features for Affect RecognitionMultimodal Affect RecognitionAvailable APIsDiscussionConclusion

 4 Concept Extraction from Natural Text for Concept Level Text AnalysisIntroductionThe patterns for concept extractionExperiments and ResultsConclusion

 5 EmoSenticSpace: Dense concept-based affective features with common-sense knowledgeIntroductionLexical Resources UsedFeatures Used for ClassificationFuzzy ClusteringHard ClusteringImplementationDirect Evaluation Of The Assigned Emotion LabelsConstruction Of EmosenticspacePerformance on ApplicationsSummary of Lexical Resources and Features UsedConclusion

 6 Sentic Patterns: Sentiment Data Flow Analysis by Means of Dynamic Linguistic PatternsIntroductionGeneral rulesCombining sentic patterns with machine learning for text-basedsentiment analysisEvaluationConclusion 

7 Combining Textual Clues with Audio-Visual Information for Multimodal Sentiment AnalysisIntroductionExtracting Features from Textual DataExtracting Features from Visual DataExtracting Features from Audio DataExperimental ResultsSpeeding up the computational time: The role of ELMImproved multimodal sentiment analysis: Deep learning-basedvisual feature extractionConvolutional Recurrent Multiple Kernel Learning (CRMKL)Experimental Results and DiscussionConclusion 

8 Conclusion and Future WorkSocial ImpactAdvantagesLimitationsFuture Work

Index