New Frontiers in Textual Data Analysis

Preface
Contents
Part I Statistical Methods for Textual Data Analysis
	Statistical and Deep Learning Methods for a Linguistic and Literary Analysis
		1 Introduction
			1.1 An Integrated Methodology
		2 Analysis
			2.1 Return to the Text. in Search of Morphosyntactic Patterns
			2.2 Aesthetic Intent
		3 Conclusion
		References
	Statistical Profiling of Hybrid CNN-SVM Effectiveness
		1 Introduction
		2 Related Work
		3 Experimental Protocol
			3.1 Datasets
				3.1.1 EMNIST
				3.1.2 Arabic Handwritten Characters Dataset (AHCD)
				3.1.3 MADBase
			3.2 Evaluation Methodology
				3.2.1 The SVM Architecture
				3.2.2 The CNN Architecture
				3.2.3 The Hybrid CNN-SVM Architecture
		4 Experimental Results
		5 Conclusion
		References
	EMOtivo: A Classifier for Emotion Detection of Italian Texts Trained on a Self-Labelled Corpus
		1 Introduction
		2 Conceptual Framework
		3 Methods
			3.1 Selecting the Most Representative Emojis for Each Emotion
			3.2 Building the Dataset
			3.3 Cleaning and Balancing
			3.4 Model
		4 Results
		5 Conclusion
		References
	Community Detection and Semantic Analysis on Twitter. The Case of “No Green Pass” and “No Vax” Movement in Italy
		1 Introduction
		2 Methodology, Network, and Lexical Framework
		3 Data Description
		4 Results
			4.1 Communities Identification
			4.2 Textual Correspondence Analysis: CA-GALT Approach
		5 Conclusions
		References
	Predicting Customer Satisfaction by Amazon Reviews: The Bluetooth Earphones Example
		1 Introduction
		2 Dataset Description
		3 Brand Insights by Latent Dirichlet Allocation
		4 Star Rating Prediction by LDA Topics
			4.1 Logistic Regression Model
			4.2 Classification Trees
		5 Final Remarks
		References
	Symmetric Non-negative Matrix Factorization for Analyzing the Scientific Production on Day Surgery
		1 Introduction
		2 Symmetric Non-negative Matrix Factorization
		3 Data and Materials
		4 Main Results
		5 Discussion and Conclusion
		References
	Social Media Effects on Sales: Consumer Sentiment in a State-Space Model
		1 Introduction
		2 Twitter Corpora
		3 VADER Sentiment Scoring
		4 ETS Benchmarks and e-WOM MARSS Model
		5 Main Results
		6 Conclusions
		References
Part II Advances in Language Processing
	Quality Enhancements in Experimental Statistics: The Italian Social Mood on Economy Index
		1 Introduction
		2 Methods
			2.1 Data-Driven Analysis Through Word Embedding
			2.2 Lexicon-Based Methods
		3 Results
			3.1 The Filter
			3.2 The Vocabulary
			3.3 Comparison with Other Time Series
		4 Conclusions
		References
	A Strategy to Identify the Peculiarity of a Lexicon in the Analysis of a Corpus
		1 Introduction
		2 The Statistical-Linguistic Resource of Standard Italian
		3 TaLTaC_4
			3.1 Computational Aspects for the Measurement of Peculiarity
		4 MWE Algorithm
		5 Application of the Strategy
			5.1 The Calculation of the Peculiarity
			5.2 Search for Significant MWE Entities
		6 Conclusion
		References
	Neologisms and Estrangement in a Corpus of Science Fiction
		1 Introduction
		2 Science Fiction and Estrangement
		3 Description of the Corpus
		4 Identification and Quantitative Analysis of Neologisms
		5 Neologism Construction Processes in the PARVIS Corpus
			5.1 Use of Non-Latin Alphabets and Loanwords
			5.2 Phonetics-Based Neologisms
			5.3 Creation of Portmanteau Words
			5.4 Unknown Words Analysable Through a Morphological Pattern
			5.5 Processes for the Neologism Production
		6 Neologism Contribution to Estrangement
		7 Conclusions and Future Work
		References
	Automatic Retrieving of Derived Quechua Verbs
		1 Introduction
		2 Fundamental Conjugation
		3 Verb to Verb Derivations
		4 Topological Transformations of the Indefinite Tense IT
		5 Multi-Suffix Derivations
		6 Mixed Conjugated Forms of IPS and PPS
		7 Retrieval of Inflected and Derived Verbs
		8 A Socio-Linguistic Application
		9 Conclusion and Perspectives
		References
	A Stylometric Profile of Carmen Mola in the Gender of `Her\' Authorial Persona and the Contribution of Each Author Behind the Pseudonym
		1 Introduction
		2 Background
		3 Corpus
		4 Methodology
		5 Results
			5.1 Gender Profiling
			5.2 Closeness of Component Authors
			5.3 Material Contributions
		6 Discussion
		7 Conclusions
		References
	Automatic Genre Classification of Czech Texts Based on Syntactic Functions
		1 Introduction
		2 Language Material
		3 Methodology
		4 Results
		5 Discussion and Conclusion
		References
	Deep Learning As an Aid to Text Mining in the Choice of Texts to Lemmatise for a Comparison Corpus: A Stylistic Study of Peter Damian\'s Letters
		1 Introduction
		2 The Methods of the LASLA: Reliable, But Time-Consuming
		3 A Methodological Confrontation: TDA and Deep Learning
			3.1 Methodology: The Constitution of the Comparison Corpus
			3.2 Period 1: Antiquity
				3.2.1 The Results of the TDA
				3.2.2 The Results of the Deep Learning
			3.3 Period 2: Late Antiquity and High Middle Ages
				3.3.1 The Results of the TDA
				3.3.2 The Results of the Deep Learning
			3.4 Period 3: Central Middle Ages
				3.4.1 The Results of the TDA
				3.4.2 The Results of the Deep Learning
		4 Evaluation
		References
	Dialogic Process Analysis in Natural Language Processing: An Attempt to Describe the Sense of Reality and Meaning of Textdata
		1 Introduction
		2 Dialogic Science Research Programme and Aim
			2.1 Dataset
			2.2 Analysis
		3 Conclusions
		References
	Multichannel Convolutional Transformer and Intertextuality: A Latin Case Study
		1 Introduction
		2 Model
			2.1 Multichannel CNN and Multi-head Attention
			2.2 Intertextuality Detection Tool
		3 Compared Approaches
			3.1 Intertextuality Detection
		4 Conclusion
		References
Part III Emotional and Sentiment Analyses
	Opinion Mining Hybrid Approach. An Application to Investigate the Users\' Political Positions in Disinformative Echo Chambers
		1 Disinformation: A Difficult Phenomenon to Define
		2 Methodology and Data Collection
		3 A Hybrid Approach Proposal to Opinion Mining
		4 The Users\' Political Orientation
		5 The Users\' Polarized Narratives in the Disinformative Echo Chambers
		6 First Conclusions and Future Work
		References
	Prediction of Italians\' Sentiment During the First COVID-19 Lockdown Through a Weighted Random Forest Balanced with SMOTE Algorithm
		1 Introduction
		2 Data
		3 Methods
			3.1 Textual Analysis
			3.2 Sentiment Prediction
		4 Results
			4.1 Differences Between Respondents and Nonrespondents in the Final Question
			4.2 Sentiment Classification
		5 Conclusion
		References
	Sentiment Analysis on Social Network Data: The Regional IndexRETI
		1 Introduction
		2 Literature Review
		3 Data
		4 Methods
		5 Results and Discussion
		6 Conclusions
		References
	Integrating Text Mining and Hermeneutic Analysis: The Case of International Volunteering Biographies
		1 Introduction
		2 Methods
		3 Results
		4 Conclusion
		References
	Emotional Text Mining and Multilingual Corpora: The Analysis of #Covid-19 on Twitter
		1 Introduction
			1.1 On the Nature of the Hashtag
		2 Emotional Text Mining of Multilingual Corpora
		3 #Covid-19 in Europe
		4 Discussion and Conclusions
		References
	Emotional Markers As Indicators of Investor Attitudes: EDA Sub-process Proposal
		1 Introduction
		2 Investors\' Psychological Mechanisms in COVID-19
		3 Pseudocode
		4 Analysis Results
		5 Conclusions
		References
	“The Spanish Model”: An Analysis of Spanish Culture in Organ Donation Through the Emotional Text Mining
		1 Introduction
		2 Method
		3 Results
			3.1 The Four Factors
			3.2 The Five Clusters
		4 Results
		References
Part IV Textual Data Analysis in Action
	Comparative Analysis of National Reports: The Case of the Erasmus+ ECOLHE Project
		1 Introduction
		2 Methods
		3 Results
		4 Discussion and Conclusions
		References
	Action Research in Psychology: The Case of Adoptive Family
		1 Introduction
		2 Theoretical Framework
		3 The Research
			3.1 Participants
			3.2 Data Analysis
			3.3 Results
			3.4 Discussion and Conclusions
		References
	Uncovering Uncertainty in Narrative Economics: A Semantic Search Approach
		1 Introduction
		2 Methodology
			2.1 Topic Models
			2.2 Central Banking Communication
		3 Data and Results
		4 Conclusion
		References
	Italian Institutional Communication in Pandemic Period: A Chronological Analysis of Prime Minister Speeches
		1 Introduction
		2 Data Selection and Preparation
		3 Time Series of New Positives and Emotional Analysis
		4 Text Classification
		5 Conclusions
		References
	What About Corruption? A Text Analytics Method for a Scoping Literature Review
		1 Introduction
		2 Data Collection and Analysis
			2.1 Corpus Built Around the Term `Corruption\'
			2.2 Corpus Built Around the Term `Transparency\'
		3 Discussion and Conclusions
		References
	The Stability of the Discursive Framework of the Ministry of Foreign Affairs Tested by Textometry
		1 Introduction
		2 Research Context: The MFA and Its Communication
		3 Corpus and Method
		4 The Construction of Discourse in the Political Arena: The Case of Statements
		5 The Weight of Statuses
		6 Conclusion
		Références
	Statistical Analysis of Textual Data for Longitudinal Analysis. A Study on Postgraduate Course Participants\' Reflections
		1 Introduction
		2 Journalism\'s Postgraduate Course: “Telling the Truth. How to Inform by Promoting an Inclusive Society”
			2.1 Critical Awareness
			2.2 Reflexivity
			2.3 The Propensity for Inclusion
		3 Statistical Analysis of Textual Data for Longitudinal Analysis
			3.1 Corpus and Method
		4 Results
		5 Conclusion
		References
	Third Mission and VQR 2015–2019: A Bigram\'s Story
		1 Introduction and Third Mission Background
		2 Data Collection
		3 Methods
		4 Results: Strategic Communication of TM (RQ1)
		5 Results: Association Between ANVUR Score and Strategic Communication of TM (RQ2)
		6 Discussion and Conclusions
		References