Text Mining

After many years of working with undergraduate, postgraduate and research students we recognise only too well the struggles that they often experience wrestling with the somewhat strange and seemingly obtuse language used to describe research philosophy. We once experienced similar difficulties and empathise with the confusion and lack of confidence that flows from being unclear whether you have really understood terms such as methodology, ontology or epistemology. We set out to produce a text that dealt with two problems. The first was to provide something that guides novice researchers through the whole process from identifying a topic to the writing up of findings via engagement with the literature and a brief overview of both qualitative and quantitative techniques. The second problem we wanted to tackle related to what we often refer to as ‘the ologies’. Here we wanted to offer a structured approach to familiarising yourself with the terminology and to demonstrate how a nested set of descriptions builds towards a coherent, comprehensive and consistent articulation of your research paradigm.

We are indebted to our colleagues for their help in delivering on the first
of these two problems in the first edition of the book. This was achieved at a pace which seemed frankly ridiculous but which produced a remarkably coherent guide for novice researchers. Despite positive feedback on many aspects of the first edition from both students and colleagues, we were however convinced that we could improve in relation to ‘the ologies’.

For this reason, the second edition features some relatively minor
changes to many chapters and a complete rewrite of our account of research philosophy. Central to the revised text is the methods map (see Chapter 4), which sets out a logical process for researchers to articulate their position in relation to five key aspects of their research philosophy. We have road tested this approach with many colleagues and students to ensure that it is clear and concise. In addition, we have developed a free app to accompany the book and this enables novice researcher to quickly develop a comprehensive justification of their particular research design in an interactive way. We would acknowledge that the methods map makes some simplifications and would suggest that for all but the most sophisticated of purposes, this is entirely appropriate. Indeed, if you are well enough versed in the philosophical nuances of knowledge explored in the method map then you are probably not part of our intended audience since you already possess the skills, confidence and capacity to articulate and defend the underpinning philosophical assumptions of your research. For everyone else, we hope that the second edition of Research Methods for Business and Management helps demystify the dreaded ‘ologies’.

In social media platforms, hate speech can be a reason of "cyber conflict" which can affect social life in both of individual-level and country-level. Hateful and antagonistic content propagated via social networks has the potential to cause harm and suffering on an individual basis and lead to social tension and disorder beyond cyber space. However, social networks cannot control all the content that users post. For this reason, there is a demand for automatic detection of hate speech. This demand particularly raises when the content is written in complex languages (e.g. Arabic). Arabic text is known with its challenges, complexity and scarcity of its resources. This paper will present a background on hate speech and its related detection approaches. In addition, the recent contributions on hate speech and its related antisocial behaviour topics will be reviewed. Finally, challenges and recommendations for the Arabic hate speech detection problem will be presented.

In recent years, there has been an exponential growth in the number of complex documents and texts that require a deeper understanding of machine learning methods to be able to accurately classify texts in many applications. Many machine learning approaches have achieved surpassing results in natural language processing. The success of these learning algorithms relies on their capacity to understand complex models and non-linear relationships within data. However, finding suitable structures, architectures, and techniques for text classification is a challenge for researchers. In this paper, a brief overview of text classification algorithms is discussed. This overview covers different text feature extractions, dimensionality reduction methods, existing algorithms and techniques, and evaluations methods. Finally, the limitations of each technique and their application in real-world problems are discussed.

A Large number of digital text information is generated every day. Effectively searching, managing and exploring the text data has become a main task. In this paper, we first represent an introduction to text mining and a probabilistic topic model Latent Dirichlet allocation. Then two experiments are proposed-Wikipedia articles and users' tweets topic modelling. The former one builds up a document topic model, aiming to a topic perspective solution on searching, exploring and recommending articles. The latter one sets up a user topic model, providing a full research and analysis over Twitter users' interest. The experiment process including data collecting, data pre-processing and model training is fully documented and commented. Further more, the conclusion and application of this paper could be a useful computation tool for social and business research.

perkembangan smartphone di Indonesia yang sangat pesat, menjadi cikal bakal pertumbuhan aplikasi mobile. penduduk Indonesia dengan jumlah penduduk sebanyak 250 juta jiwa adalah pasar yang besar, sehingga banyak peluang yang dimanfaatkan oleh para kalangan, salah satunya adalah kebutuhan transportasi online yang dapat memudahkan masyarakat dalam pemesanan transportasi secara online, salah satunya adalah aplikasi Go-jek, Antusiasme masyarakat dalam menyambut platform ini begitu besar, terbukti dengan kedudukan Go-jek pada rating tertinggi yang dipilih konsumen sebanyak 72,6 persen. Banyaknya komentar masyarakat yang diberikan kepada aplikasi ini merupakan hal yang menarik untuk diolah menjadi suatu informasi. Pemanfaatan informasi tersebut membutuhkan teknik analisis sehingga informasi yang dihasilkan dapat membantu banyak pihak. Metode yang digunakan pada penelitian ini adalah metode klasifikasi Naïve Bayes. Naïve Bayes Classifier merupakan teknik pembelajaran algoritma data mining yang memanfaatkan metode probabilitas dan statistic. Metode ini gunakan untuk mengklasifikasikan komentar yang diberikan masyarkat terhadap aplikasi Go-jek. Komentar yang diberikan pada apilkasi ini nantinya akan di klasifikasikan menjadi komentar postifi, negatif, dan netral. Tujuan dilakukannya penelitian ini adalah untuk mengetahui seberapa besar komentar positif, negative dan netral yang diberikan masyarakat terhadap aplikasi ini, sehingga nantinya akan memudahkan pihak Go-jek dalam memberikan kebijakan atau pengembangan aplikasi di masa depan. (paper ini belum di published karena masih banyak kekurangan hanya ingin berbagi melalui academia ini)

The continually increasing number of complex datasets each year necessitates ever improving machine learning methods for robust and accurate categorization of these data. This paper introduces Random Multimodel Deep Learning (RMDL): a new ensemble, deep learning approach for classification. Deep learning models have achieved state-of-the-art results across many domains. RMDL solves the problem of finding the best deep learning structure and architecture while simultaneously improving robustness and accuracy through ensembles of deep learning architectures. RDML can accept as input a variety data to include text, video, images, and symbolic. This paper describes RMDL and shows test results for image and text data including MNIST, CIFAR-10, WOS, Reuters, IMDB, and 20newsgroup. These test results show that RDML produces consistently better performance than standard methods over a broad range of data types and classification problems.

Artificial intelligence (AI) has been applied widely in our daily lives in a variety of ways with numerous successful stories. AI has also contributed to dealing with the coronavirus disease (COVID-19) pandemic, which has been happening around the globe. This paper presents a survey of AI methods being used in various applications in the fight against the COVID-19 outbreak and outlines the crucial roles of AI research in this unprecedented battle. We touch on a number of areas where AI plays as an essential component, from medical image processing, data analytics, text mining and natural language processing, the Internet of Things, to computational biology and medicine. A summary of COVID-19 related data sources that are available for research purposes is also presented. Research directions on exploring the potentials of AI and enhancing its capabilities and power in the battle are thoroughly discussed. We highlight 13 groups of problems related to the COVID-19 pandemic and point out promising AI methods and tools that can be used to solve those problems. It is envisaged that this study will provide AI researchers and the wider community an overview of the current status of AI applications and motivate researchers in harnessing AI potentials in the fight against COVID-19.

Citation/Export
MLA

Miss. Ankita R. Makode, Prof. Arvind S. Kapse, “Knowledge Discovery using Various Multimedia Data Mining Technique”, March 15 Volume 3 Issue 3 , International Journal on Recent and Innovation Trends in Computing and Communication (IJRITCC), ISSN: 2321-8169, PP: 1138 - 1141, DOI: 10.17762/ijritcc2321-8169.150354
APA

Miss. Ankita R. Makode, Prof. Arvind S. Kapse, March 15 Volume 3 Issue 3, “Knowledge Discovery using Various Multimedia Data Mining Technique”, International Journal on Recent and Innovation Trends in Computing and Communication (IJRITCC), ISSN: 2321-8169, PP: 1138 - 1141, DOI: 10.17762/ijritcc2321-8169.150354

La producción y comercialización de arte textil es una alternativa importante para la generación de ingresos para alrededor de 12 000 artesanas y artesanos en situación de pobreza de las comunidades de la sierra rural del Perú y Bolivia. Si bien los/as más pobres son los que conservan los aspectos culturales del arte textil local, el potencial del arte textil no es aprovechado debido a una serie de factores como la producción de arte textil a artesanía de mala calidad, bajos niveles de comercialización, ventas como productos de segunda categoría, discriminación, exclusión social y cultural, entre otros motivos. Existen serias limitaciones técnicas, comerciales y de gestión que obstaculizan el desarrollo y consolidación de la cadena de valor. En la actualidad la producción de arte genera bajos ingresos, en unas cadenas de valor poco equitativa en la distribución de la renta para las familias involucradas.

En este marco se plantea la elaboración de un inventario de tecnologías, técnicas, iconografía, uso de color, significados, principales soportes y conocimientos tradicionales vinculados a la producción de arte textil bajo el enfoque de que estos constituyen fuente de información, entendiendo que el textil fue y es un soporte para la transmisión de conocimientos en los Andes. El documento que presentamos a continuación, es fruto de este trabajo, que se realizó de forma participativa y a través de visitas a las comunidades y familias dedicadas a la artesanía textil, quienes expusieron desde su propia mirada y vivencia los usos y características del arte textil tradicional.

Electronic Theses and Dissertations (ETDs) poses the challenge of managing and extraction of appropriate knowledge for decision making. To tackle the same, topic modeling was first applied to Library and Information Science (LIS) theses submitted to Shodhganga (an Indian ETDs digital repository) to determine the five core topics/tags and then the performance of the built model based on those topics/tags were analyzed. Using a Latent Dirichlet Allocation based Topic-Modeling-Toolkit, the five core topics were found to be information literacy, user studies, scientometrics, library resources and library services for the epoch 2013-2017 and consequently all the theses were summarized with the presence of their respective topic proportion for the tags/topics. A Support Vector Machine (Linear) prediction model using RapidMiner toolbox was created and showed 88.78% accuracy with 0.85 kappa value.

Purpose – This paper examines supply chain risk management (SCRM) from a holistic systems thinking perspective by considering the different typologies that have evolved as a result of earlier research. The purpose of this paper is to identify important strategic changes in the field and to outline future requirements and research opportunities in SCRM.

Design/methodology/approach – The systematic literature review (SLR) methodology employed by this research was used to evaluate and categorise a literature survey of quality articles published over a period of ten years (2000-2010). Additionally, the findings from the SLR have been strengthened through cross validation against results obtained from an associated text mining activity.

Findings – The SLR methodology has provided a rich, unbiased and holistic picture of the advances in the field of SCRM. Consequently, important new research areas have been identified based on a multi-perspective descriptive and thematic data analysis. In addition, the analysis, based on evolved typologies, indicates a growth of SCRM from a nascent to a fairly established activity over the past decade.

Practical implications – The systematic approach undertaken for the literature review will provide future researchers and managers with an insightful understanding of the scope of the SCRM field. Also, the literature review provides important clues on new research directions for SCRM through identification of gaps in current knowledge.

Originality/value – The holistic approach to SCRM was found to be an important missing link in earlier literature surveys. The outcome of the SLR reported in this paper has provided critical insights into the present and future scope of the SCRM field. The identified research insights, gaps and future directions will encourage new research techniques, with a view to managing the risks in the globalized supply chain environment."""

The literature review on random forests and text mining provided in this paper makes clear the link and relevance that exists between these two fields, and shows how academia and industry are doing an increasing number of studies on these subjects, keeping them an interesting issue for further development.

The growing world of information technology course, the growing impact of data outstanding and continues to grow significantly, and initial data processing or preprocessing text in text mining process is expected to reduce by eliminating the word-the word or text that are not necessary or do not have the meaning of text database or document. By decreasing the amount of text was expected to ease further processing in order to mine the information contained within the document-document or text-text in a miraculous process by applying existing methods to produce useful information from the text without reducing the sense or meaning and information contained in the document.

Ease of information exchange is frequently abused by some people. The abuse happens on academic envirotment and is commited by students. One of abuse act is plagiarism on process while completing subject, practicum and final assignment. It is one of plagiarisme that commited by students. Lecturer's method to anticipate plagiarisme act is to compare students's assignment independently. Lecturers have difficulty on evaluation of learning outcomes because number of subject participant is many. Method for comparing two objects is required to designing application for detecting similarity on text document. Cosine Similarity is one of method that can be implemented to compare text documents. Cosine Similary uses two vector parameters. The vector is based on number of similar word on both text documents so that Cosine Similarity is not be able to recognize meaning of sentence. Nazief & Adriani algorithm is algorithm to transform affix word into root word in Indonesia language. Comparing words are transformed into root word firstly in order to make Cosine Similarity's calculation more optimal. Final assignment results similarity of detection application on text document that implementing Nazief & Adriani algorithm and Cosine Similarity method. 1. PENDAHULUAN 1.1. Latar Belakang Dewasa ini, perkembangan teknologi informasi berkembang dengan pesat dan memberikan dampak positif. Salah satu dampak positif perkembangan teknologi informasi adalah kemudahan dalam bertukar informasi. Kemudahan tersebut sering disalahgunakan oleh seseorang atau beberapa orang dalam penyelesaian pekerjaan. Penyalahgunaan tersebut merupakan salah satu dampak negatif dari perkembangan teknolog informasi. Penyalahgunaan ini sering terjadi di lingkungan akademis terutama oleh mahasiswa. Salah satu tindakan penyalahgunaan yang terjadi adalah melakukan penjiplakan terhadap seluruh atau beberapa teks dari satu atau beberapa berkas dokumen teks sumber ke berkas dokumen teks lain. Praktis penyalahgunaan ini sering terjadi pada proses pembuatan tugas mata kuliah, tugas praktikum dan tugas akhir. Plagiat adalah pengambilan karangan (pendapat) orang lain dan disiarkan sebagai karangan (pendapat) sendiri [1]. Tindakan pengambilan karangan atau ide bukan suatu hal yang baru bagi mahasiswa. Adanya alat bantu seperti komputer dan laptop, memberikan kemudahan untuk melakukan tindakan penjiplakan. Hal tersebut berdampak pada perilaku yang terkesan malas dalam mengerjakan tugas mata kuliah, tugas praktikum dan tugas akhir. Hal tersebut juga berpengaruh pada tahap evaluasi hasil pembelajaran. Peserta mata kuliah yang tidak sedikit memberikan kesulitan bagi para pengajar untuk memberikan hasil yang objektif. Hal tersebut menimbulkan kecurigaan terhadap tugas-tugas yang nantinya akan diakamulasikan ke tahap nilai akhir. Sehingga para pengajar akan direpotkan untuk menganalisis satu per satu tugas mahasiswa dikarenakan jumlah peserta mata kuliah tidak sedikit. Beberapa pengajar masih menggunakan cara manual untuk menentukan kemiripan suatu dokumen teks. Cara yang dimaksud adalah membandingkan dua atau lebih berkas dokumen dan menelusuri paragraf per paragraf dan menentukan kata, kalimat atau paragraf yang memiliki kemiripan. Cara tersebut

The construction of computer models is becoming an increasingly useful and popular way of testing theories in the cognitive sciences. This paper will present a brief overview of the methods available for constructing and testing computer models of social phenomena such as religious beliefs and behaviors. It will focus on the importance of theoretical continuity and data replication in computer modelling while negotiating the relationship between specificity and ecological validity when models are extended into novel contexts. This paper will argue that computer modeling is an important supplement to the methodological toolbox of cognitive scientists interested in human social phenomena. However, this is only the case if developers pay close attention to research methods and theories and if the method of a model’s development is appropriate for the target phenomenon (Sun, 2006). It concludes that multi-agent AI models are the most appropriate computational tool for the study of complex social phenomena.

The continually increasing number of documents produced each year necessitates ever improving information processing methods for searching, retrieving, and organizing text. Central to these information processing methods is document classification, which has become an important application for supervised learning. Recently the performance of these traditional classifiers has degraded as the number of documents has increased. This is because along with this growth in the number of documents has come an increase in the number of categories. This paper approaches this problem differently from current document classification methods that view the problem as multi-class classification. Instead we perform hierarchical classification using an approach we call Hierarchical Deep Learning for Text classification (HDLTex). HDLTex employs stacks of deep learning architectures to provide specialized understanding at each level of the document hierarchy.

— The present paper was given birth by the facts that political speeches emerged in Classical Rhetoric times as the primary object of linguistic interest, and, second, political speeches have preserved throughout the centuries both their form and their considerable manipulative potential. A possible explanation for both those facts could be the 'simplicity' of political speeches. But how does one account for or give a definition of 'simplicity'? The paper aims to present a cognition-based approach the structure of political speeches in particular. It also demonstrates how simple structures deriving from basic bodily experience can serve manipulative purposes.

In this paper we use Twitter data to assess customers early reactions to the launch of two new products by Apple and Samsung by analyzing the streams generated in a 72 h window around the two events. We present a methodology based on conversational analysis to extract concept maps from Twitter streams and use semantic and topological metrics to compare the conversations. Our findings show that there are significant differences in the structural patterns of the two conversations and that the analysis of these differences can be highly informative about early customers perceptions and value judgments associated with the competing products

The advent of digital humanities has brought the notion of "big data" into the purview of humanistic inquiry. Humanists now have access to huge corpora that open research possibilities that were unthinkable a decade or two ago. However, working with corpora requires a rather different approach that is more characteristic of sciences than humanities. Namely, one has to be transparent and explicit with regard to how data are extracted and how they are analyzed. Text-mining techniques rely on explicit algorithms because they help tracing mistakes, correcting them and, ultimately, improving results. Analytical procedures for studying extracted data rest on explicit algorithms for the same reason. As a way of constructing algorithms, modeling is part and parcel of developing complex computational procedures. ...

(The webcast of the talk is available at the URL below (paper timestamp: 2:48:00; discussion timestamp: 3:51:30))

In today's digital world, food app like Zomato is widely used because it provides a platform for people to share their opinion about the restaurants and cafes they have visited. This research paper includes analysis of client ratings and reviews in Zomato utilizing content mining. Utilizing content mining, break down the content audits/reviews from the client with a specific end goal to create productive result and legit surveys. Gathering surveys dataset and handling it to check the reliability of the rating given and audit composed by client. Ascertaining reliability of the eatery subsequent to dissecting the surveys as indicated by the administration gave and cost estimation. All through this procedure look at the client audit premise on their content setting and it demonstrates that how they feel about their visit to that place.

This study explores key content and themes from online reviews to explain major service attributes of peer-to-peer (P2P) accommodation sought by guests. The results from lexical analyses indicate that attributes frequently mentioned in guest reviews are associated with location (proximity to point of interest and characteristics of neighborhood), host (service and hospitality), and property (facilities and atmosphere). Reviews focusing on location and feeling welcome are consistently linked with higher rating scores, including accuracy, cleanliness, check-in, communication, value, and overall ratings. This confirms that P2P accommodation appeal to consumers who are driven by experiential and social motivations. Marketing implications are provided.

Our focus is on methods of cultural analysis, and specifically on those methods that are formal in the sense that they rely upon the purposeful gathering (or simulating) of cultural data and a systematic analysis that involves at least some mathematically based technique. The meaning of culture is more complex, as scholars have understood cultural phenomena in very different ways across the decades and across the disciplines.  These different goals in cultural analysis give rise to different types or styles of formal methods of cultural analysis. We briefly review the history of these methods in the social sciences and finish with a description of  some of the main arenas within which formal methods of cultural analysis are being deployed and actively developed today.

"In the digital world, which is currently evolving and changing at such a rapid pace, the security of digital
information has become increasingly more important. Great development of secret contact and communication in
the world, the necessity of communication security is becoming more and more significant. Cryptography has
specific role to protect secret files like secret documents from unauthorized access. Cryptography algorithms are
classified in to two types, Public-key producing and symmetric-key producing algorithms. In this paper, I suggested
a new robust cryptography algorithm to increase security in the Symmetric-key producing algorithm. The main
features of cryptography algorithm defined in this article are the ability to encrypt the secret file in successive stages,
changing the physical structure of the secret file, no limitation for the number of keys, Creating five keys at each
stage of cryptography, storing a part of secret file at one of the keys at each stage of cryptography, Interdependence
of all keys in all stages of encrypting and decrypting, To make the keys interdependent and to encrypt the secret file
by each of them, there are 2 independent algorithms to select the type of algorithm needed to make the keys
interdependent by the user, bigger changes in the physical structure of the encrypted file In case of wrong decryption
and to make the resulting keys and encrypted file unique after the cryptography process."

Avec l'accroissement de la quantité d'information disponible sur Internet et suite au développement fulgurant de l'informatique, des besoins réels se posent maintenant au niveau de l'extraction de l'information pertinente noyée dans des giga-octets voire des téraoctets de données textuelles véhiculées dans les différents supports, En effet, au vu du flot d'information que nous connaissons; accéder aujourd'hui à l'information textuelle utile est devenu un vrai « casse-tête », il devient de plus en plus important d'aider les utilisateurs à accéder plus rapidement à l'information recherchée et à développer de nouveaux outils de recherche d'information. Des milliers de documents sont disponibles facilement à travers les réseaux et les supports informatiques dont plus de 90% sont des documents textuels, il devient désormais difficile d'accéder aux informations sans l'aide d'outils spécifique. [1] Devant les besoins naissant provoqués par cette réalité préoccupante, Le Text Mining ou fouille de données textuelles qui est une spécialisation du Data minig répond, en parti, à cette problématique. L'étude du text mining repose particulièrement sur des liens très étroits entre des recherches en linguistique textuelle et sur les formalisations adéquates en vue d'une réalisation informatique. Le text mining est traditionnellement connu pour sa faculté à extraire les concepts clés de sources d'information texte. Il propose une approche méthodologique pour exploiter l'information non structurée et la combiner avec les techniques de data mining usuelles. Il permet de « comprendre» (mining) artificiellement le texte (text) en vue d'extraire une quantité d'information limitée mais pertinente afin de répondre au besoin recherché qui est principalement le gain de temps pour une meilleure productivité et parmi ces application connue on note plus spécialement la classification automatique de textes. On appelle classification automatique la catégorisation algorithmique d'objets. Celle-ci consiste à attribuer une classe ou catégorie à chaque objet (ou individu) à classer. Elle peut être : supervisée dont les classes sont connues à priori, elles ont en général une sémantique associée ou non-supervisée (en anglais clustering) où les classes sont fondées sur la structure des objets, les classes possibles ne sont pas reconnu à l'avance le but pour les construire est de regrouper dans un même cluster les objets considérés comme similaires.
Dans les deux cas, on a besoin de définir la notion de distance entre deux classes: le critère d'agrégation. Le cadre général dans lequel s'inscrit ce mémoire est celui de la classification automatique non-supervisée dans laquelle aucune hypothèse sur les classes n'est connue, on dispose d'un ensemble de textes, issus de différents domaines, et on essaye de trouver une liaison fonctionnelle entre ces textes. Notre choix est basé sur l'utilisation de la méthode de clustering spectral pour la classification non supervisée des documents texte et pour leur représentation nous avons choisi la méthode de sac de mots (ou bag-of-words en anglais). Notre manuscrit est structuré en cinq chapitres. Les quatre premiers sont à caractère fondamental et traitent les différents aspects de la classification spectrale. Le dernier est applicatif et il est dédié à la partie implémentation. Dans le premier chapitre, nous présenterons le Data Mining qui s'intéresse à l'extraction des connaissances à partir des données en faisant aussi appel à des techniques généralement statistiques. Dans le deuxième chapitre nous exposerons la notion du text mining ainsi que ses objectifs et ses taches. Dans le troisième, vu que pour traiter un document, tous les algorithmes ne travaillent pas nécessairement directement sur le document lui même, nous expliquerons les différentes techniques de représentation des textes. Le quatrième chapitre est divisé en deux parties, la première est consacrée à la présentation des modèles d'apprentissage non supervisé. Il présente le contexte de l'étude de la classification, en expliquant en détail différentes approches de classification non supervisée. Dans le cadre de ce mémoire, nous nous sommes intéressés plus particulièrement aux algorithmes à base de la classification non hiérarchique (par partitionnement). La deuxième partie présente un état de l’art sur les algorithmes de classification spectrale basés sur une bi-partition ou une K-partition des données. Le principe est de définir un espace de projection spectral à partir des vecteurs propres de la matrice Laplacienne et de rechercher dans cet espace des groupes représentatifs des données. Les algorithmes spectraux dépendent des paramètres de la fonction similarité et du nombre de groupes recherchés. Et enfin dans le dernier chapitre nous mettrons en œuvre, l'approche que nous proposons, avec une description détaillée des étapes de l'implémentation. Une conclusion générale vient pour clôturer ce travail.