DOCSLIB.ORG

Analysis of Methods and Means of Text Mining Z

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Analysis of Methods and Means of Text Mining Z ECONTECHMOD. AN INTERNATIONAL QUARTERLY JOURNAL – 2017. Vol. 6. No. 2. 73–78 Analysis of methods and means of text mining Z. Rybchak 1, O. Basystiuk 2 1Lviv Polytechnic National University, e-mail: [email protected] 2Lviv Polytechnic National University, e-mail: [email protected] Received February 15.2017: accepted May 28.2017 Abstract. In Big Data era when data volume doubled every business rules, and relationships – that is otherwise locked year analyzing of all this data become really complicated task, in textual form, impenetrable to automated processing. so in this case text mining systems, techniques and tools become main instrument of analyzing tones and tones of information, As said earlier 70% of business-relevant information selecting that information that suit the best for your needs and is stored as text, but this is truism for all spheres of human just help save your time for more interesting thing. The main life, most information is currently stored as text aims of this article are explain basic principles of this field and information, that why text mining is believed to have a overview some interesting technologies that nowadays are high commercial potential value. Increasing interest is to widely used in text mining. multilingual data mining: the ability to gain information Key words: text mining, text analytics, data analysing, across languages and cluster similar items from different high-quality information, text categorization, text clustering, linguistic sources according to their meaning. document summarization, sentiment analysis. A simple application is to scan a text written in a INTRODUCTION natural language, then identify key-phrases of this document, show this key-phrases and by this phrases Text mining, also known as intelligent text analysis, program can make prediction of what type of text it is. text data mining or knowledge-discovery in text (KDT) all this terms describes a set of linguistic, statistical, and Typical tasks in text mining include this: text machine learning techniques that refers generally to the categorization, and text clustering (this tasks is main goal process of extracting interesting and non-trivial of my example of simple application), also it could be information and knowledge from unstructured text. concept extraction, entity extraction, sentiment analysis Techniques that help model and structure the information (popular, when you want to know, what people think content of textual sources for business intelligence, about your business), document summarization etc. exploratory data analysis, research, or investigation. This technology nowadays is widely implemented Text mining is a relatively young field, manual text and used in variety of government, research, and business mining researches started in the middle of 1980s, notably needs. for sciences and government needs, but interested people TEXT MINING AND MINING TECHNICS and technological advances have enabled developing of OF RECEIVING HIGH-QUALITY INFORMATION that field during this time. Nowadays this field united such field of science, like: computational linguistics, Text mining, it`s process of mining text data, or in machine learning and some specific field of statistics. other words receiving high-quality information from text The field of text mining usually deals with text whose data. High-quality text information usually refers to some main function is providing communication and help people combination of relevance, novelty, and interestingness. To to express their thoughts and opinions, and the motivation for get high-quality information from text, use few methods, trying to extract information from such text automatically is such as: information retrieval, lexical analysis to study compelling – even if success is only partial. word frequency distributions, pattern recognition, pattern The term text mining also describes that application learning, regularities in data, tagging/annotation, of text mining to respond to business problems, whether information extraction, data mining techniques including independently or in conjunction with query and analysis link and association analysis, visualization, and predictive of fielded, numerical data. It is a truism that more than analytics. In addition, text mining involves the process of 70 % of business-relevant information is stored in structuring the input text, deriving patterns within the unstructured form, such as text. These techniques and structured data, and finally evaluation and interpretation processes help to discover and present knowledge – facts, of the output. 74 Z. RYBCHAK, O. BASYSTIUK The main goal is, to turn text into data for analysis, Open source: by applications of natural language processing and · Natural Language Toolkit; analytical methods, and when you get this data for · OpenNLP; analysis, you can create method of processing this data in · Orange. way you need, or you already created libraries for this. One particular framework and development All this method is a branch of machine learning (or environment for text mining, called General Architecture nearly synonymous with machine learning), especially for Text Engineering or GATE, aims to help users recognition of patterns and regularities in data, because develop, evaluate and deploy systems for what the authors pattern recognition systems are in many cases trained term “language engineering.” It provides support not just from labeled “training" data, but when no labeled data are for standard text mining applications such as information available other algorithms can be used to discover extraction, but also for tasks such as building and previously unknown patterns. Same thing for regularities, annotating corpora, and evaluating the applications. in this case knowledge of computational linguistics is At the lowest level, GATE supports a variety of needed, you input labeled data of regularities and trained formats including XML, RTF, HTML, SGML, email and your algorithm to recognize this regularities and handle plain text, converting them into a single unified model them in right order. that also supports annotation. There are three storage TECHNIQUES AND TOOLS ANALYSIS mechanisms: · relational database; Text mining systems use a big spectrum of different · serialized Java object; approaches, partly because of the great scope of systems · XML based internal format documents can be re- and tools that perform text mining, and partly because the exported into their original format with or without field don`t have it dominant methodologies due to the annotations. youth of this field. Nevertheless, we can divide these Text encoding is based on Unicode to provide approaches on: support for multilingual data processing, so that systems · High-level, that’s mean you involved into your developed with GATE can be ported to new languages text mining application systems that use an automatic with no additional overhead apart from the development training systems to do stuff like recognising patterns. This of the resources needed for the specific language. GATE works in easy way, you create some training system, push includes a tokenizer and a sentence splitter. It incorporates into this system already target “training” data and by this a part-of-speech tagger and a gazetteer that includes lists data your system learn how to recognize different patterns. of cities, organizations, days of the week, etc. It has a · Low-level, that`s mean you deal with natural semantic tagger that applies hand-crafted rules written in language and involve your custom created decisions a language in which patterns can be described and making system, that strongly influence the success of annotations created as a result. mining process. Patterns can be specified by giving a particular text But as low-level way of solving problems is more string, or annotations that have previously been created by complexity, time-consuming and required big amount of modules such as the tokenizer, gazetteer, or document knowledge, it`s less popular nowadays. Because you need format analysis. It also includes semantic modules that to implement by yourself this trivial on first view, but recognize relations between entities and detect co- really important logic on dealing with all small decisions, reference. It contains tools for creating new language like: how to deal with apostrophes and hyphens, resources, and for evaluating the performance of text capitalization, punctuation, numbers, alphanumeric mining systems developed with GATE. strings, whether the amount of white space is significant, One application of GATE is a system for entity whether to impose a maximum length on tokens, what to extraction of names that is capable of processing texts from do with non-printing characters, and so on. So using of widely different domains and genres. This has been used to this method require good background in text analysis perform recognition and tracking tasks of named, nominal principles and implementing all logic of this small, but and pronominal entities in several types of text.Least but really important, decisions. not last, there are few popular techniques, which will help In this case, high-level way provides big you to start work in effectively mining text data. opportunities to create some small, but pretty useful app · Sentiment analysis – analyzing the people in a few minutes. There are a lot of text mining computer opinions and tone of feedback, posts, articles about programs available from many commercial and
Load more

Recommended publications
  • Text Analytics 2014: User Perspectives on Solutions and Providers
    Text Analytics 2014: User Perspectives on Solutions and Providers Seth Grimes A market study sponsored by Published July 9, 2014, © Alta Plana Corporation Text Analytics 2014: User Perspectives on Solutions and Providers Table of Contents Executive Summary .............................................................................................................. 3 Growth Drivers ...................................................................................................................................... 4 Key Study Findings ................................................................................................................................ 5 About the Study and this Report ......................................................................................................... 6 Text Analytics Basics ............................................................................................................ 7 Patterns ................................................................................................................................................. 7 Structure ............................................................................................................................................... 7 Metadata ............................................................................................................................................... 8 Beyond Text .........................................................................................................................................
    [Show full text]
  • Conceptnet 5.5: an Open Multilingual Graph of General Knowledge
    Proceedings of the Thirty-First AAAI Conference on Artificial Intelligence (AAAI-17) ConceptNet 5.5: An Open Multilingual Graph of General Knowledge Robyn Speer Joshua Chin Catherine Havasi Luminoso Technologies, Inc. Union College Luminoso Technologies, Inc. 675 Massachusetts Avenue 807 Union St. 675 Massachusetts Avenue Cambridge, MA 02139 Schenectady, NY 12308 Cambridge, MA 02139 Abstract In this paper, we will concisely represent assertions such Machine learning about language can be improved by sup- as the above as triples of their start node, relation label, and plying it with specific knowledge and sources of external in- end node: the assertion that “a dog has a tail” can be repre- formation. We present here a new version of the linked open sented as (dog, HasA, tail). data resource ConceptNet that is particularly well suited to ConceptNet also represents links between knowledge re- be used with modern NLP techniques such as word embed- sources. In addition to its own knowledge about the English dings. term astronomy, for example, ConceptNet contains links to ConceptNet is a knowledge graph that connects words and URLs that define astronomy in WordNet, Wiktionary, Open- phrases of natural language with labeled edges. Its knowl- Cyc, and DBPedia. edge is collected from many sources that include expert- The graph-structured knowledge in ConceptNet can be created resources, crowd-sourcing, and games with a pur- particularly useful to NLP learning algorithms, particularly pose. It is designed to represent the general knowledge in- those based on word embeddings, such as (Mikolov et al. volved in understanding language, improving natural lan- 2013). We can use ConceptNet to build semantic spaces that guage applications by allowing the application to better un- are more effective than distributional semantics alone.
    [Show full text]
  • Text Analytics 2014: User Perspectives on Solutions and Providers
    Text Analytics 2014: User Perspectives on Solutions and Providers Seth Grimes A market study sponsored by Published July 9, 2014, © Alta Plana Corporation Text Analytics 2014: User Perspectives on Solutions and Providers Table of Contents Executive Summary .............................................................................................................. 3 The Survey ............................................................................................................................................................. 5 Key Study Findings ................................................................................................................................................ 5 About the Study and this Report .......................................................................................................................... 6 Text Analytics Basics .............................................................................................................7 Patterns ................................................................................................................................................................. 7 Structure ................................................................................................................................................................ 7 Metadata ............................................................................................................................................................... 8 Beyond Text ..........................................................................................................................................................
    [Show full text]
  • Detecting Semantic Difference: a New Model Based on Knowledge and Collocational Association
    DETECTING SEMANTIC DIFFERENCE: A NEW MODEL BASED ON KNOWLEDGE AND COLLOCATIONAL ASSOCIATION Shiva Taslimipoor1, Gloria Corpas2, and Omid Rohanian1 1Research Group in Computational Linguistics, University of Wolverhampton, UK 2Research Group in Computational Linguistics, University of Wolverhampton, UK and University of Malaga, Spain {shiva.taslimi, omid.rohanian}@wlv.ac.uk [email protected] Abstract Semantic discrimination among concepts is a daily exercise for humans when using natural languages. For example, given the words, airplane and car, the word flying can easily be thought and used as an attribute to differentiate them. In this study, we propose a novel automatic approach to detect whether an attribute word represents the difference between two given words. We exploit a combination of knowledge-based and co- occurrence features (collocations) to capture the semantic difference between two words in relation to an attribute. The features are scores that are defined for each pair of words and an attribute, based on association measures, n-gram counts, word similarity, and Concept-Net relations. Based on these features we designed a system that run several experiments on a SemEval-2018 dataset. The experimental results indicate that the proposed model performs better, or at least comparable with, other systems evaluated on the same data for this task. Keywords: semantic difference · collocation · association measures · n-gram counts · word2vec · Concept-Net relations · semantic modelling. 1. INTRODUCTION Semantic modelling in natural language processing requires attending to both semantic similarity and difference. While similarity is well-researched in the community (Mihalcea & Hassan, 2017), the ability of systems in discriminating between words is an under-explored area (Krebs et al., 2018).
    [Show full text]
  • Text Analytics for Marketing and Customer Experience
    Cold, hard numbers and bright insights from customer comments: Text analytics for marketing and customer experience. Why text analytics? Why now? Introduction When customers are freed to express their thoughts Text analytics has been around for a long time, but in their own words, they reveal more about their only recently has it become automated. In the diagram motivations: they give marketing and customer insight below, we show the broad categories of different teams the “Why” to structured data’s “What?” In the last methods. Manual coding is still taking place, of 3 years, text analytics has successfully made the leap customer satisfaction surveys for example, but it is time from university supercomputer to the office desktop consuming and its accuracy depends on the day-to- and it’s now possible to analyse 10,000 or a million day operations of the team. Automated methods are customers’ comments with full statistical reliability in quicker, more reliable and are now widely preferred. near real-time, and while maintaining enough granularity Text analytics can analyse customer comments from to provide qualitative insights. any source, including CSat surveys, social media, email In this paper, we give an overview of the market and contact centre transcriptions. A recent IBM survey today, we share knowledge from a selection of found that predictive analytics now feature most highly consultants’ work with multi-national brands over on CMOs wish list, with more than half already involved the last 3 years, and we offer advice on how to in using analytics to capture customer insight. 1 avoid expensive dead-ends.
    [Show full text]
  • A Data-Driven Text Mining and Semantic Network Analysis for Design Information Retrieval
    A Data-Driven Text Mining and Semantic Network Analysis for Design Information Retrieval By Feng Shi Imperial College London, Dyson School of Design Engineering A thesis submitted for the degree of Doctor of Philosophy July 2018 1 Abstract Data-Driven Design is an emerging area with the advent of big-data tools. Massive information stored in electronic and digital forms on the internet provides potential opportunities for knowledge discovery in the fields of design and engineering. The aim of the research reported in this thesis is to facilitate the design information retrieval process based on large-scale electronic data through the use of text mining and semantic network techniques. We have proposed a data-driven pipeline for design information retrieval including four elements, from data acquisition, text mining, semantic network analysis, to data visualisation and user interaction. Web crawling techniques are applied to fetch massive online textual data in data acquisition process. The use of text mining enables the transformation of data from unstructured raw texts into a structured semantic network. A retrieval analysis framework is proposed based on the constructed semantic network to retrieve relevant design information and provoke design innovation. Finally, a web-based platform B-Link has been developed to enable user to visualise the semantic network and interact with it through the proposed retrieval analysis framework. Seven case studies were conducted throughout the thesis to investigate the effectiveness and gain insights for each element of the pipeline. Thousands of design post news items and millions of engineering and design peer reviewed papers can be efficiently captured by web crawling techniques.
    [Show full text]
  • NEW-YORK-2017-BROCHURE.Pdf
    THE World’s NUMBER ONE AI EVENT FOR BUSINESS NEW YORK LONDON | HONG KONG | SAN FRANCISCO ZURICH | BOSTON | SINGAPORE | CAPE TOWN 2017 SPONSORS AND PARTNERS The INDUStry PartnerS AI Summit® ARE YOU READY TO LEAD THE AI REVOLUTION? When AIBusiness.com launched as the world’s first AI news portal back in 2014, Artificial DIAMOND SPONSOR Intelligence (AI) was not yet immediately associated with the business world. Fast forward to 2017 and AI is dominating conversations as corporate leaders realise its true potential. Most Fortune 1000 organisations have now embarked on an exciting journey to a business world of unprecedented efficiencies. PLATINUM SPONSORS In 2016, The AI Summit was the first conference globally dedicated to the impact of Artificial Intelligence in business. The event is now part of the only truly global AI event series, with shows in Cape Town, London, Singapore, Boston, San Francisco, New York, Hong Kong and Zurich. The AI Summit New York, showcases America’s pioneers in AI - both the corporates that are putting AI to work every day and their solution provider partners that help them on their transformation journey. Over 100+ speakers will be sharing unique, strategic, technical insights and know-how; every session offers a plurality of views, examples, and tools for maximising the opportunity of AI. Our interactive exhibition features 45+ of the most innovative solution providers showcasing everything from chat-bots and ML applications, to AR/VR. Major tech giants are joined by enthusiastic start-ups, all in the race to create the next ground-breaking AI product GOLD SPONSOR to transform the world of work.
    [Show full text]
  • Learning Multilingual Word Embeddings in Latent Metric Space: a Geometric Approach
    Learning Multilingual Word Embeddings in Latent Metric Space: A Geometric Approach Pratik Jawanpuria1, Arjun Balgovind2∗, Anoop Kunchukuttan1, Bamdev Mishra1 1Microsoft, India 2IIT Madras, India 1fpratik.jawanpuria,ankunchu,[email protected] [email protected] Abstract different languages are mapped close to each other in a common embedding space. Hence, We propose a novel geometric approach for they are useful for joint/transfer learning and learning bilingual mappings given monolin- sharing annotated data across languages in gual embeddings and a bilingual dictionary. Our approach decouples the source-to-target different NLP applications such as machine language transformation into (a) language- translation (Gu et al., 2018), building bilingual specific rotations on the original embeddings dictionaries (Mikolov et al., 2013b), mining par- to align them in a common, latent space, and allel corpora (Conneau et al., 2018), text clas- (b) a language-independent similarity met- sification (Klementiev et al., 2012), sentiment ric in this common space to better model analysis (Zhou et al., 2015), and dependency the similarity between the embeddings. Over- parsing (Ammar et al., 2016). all, we pose the bilingual mapping prob- lem as a classification problem on smooth Mikolov et al. (2013b) empirically show that a Riemannian manifolds. Empirically, our ap- linear transformation of embeddings from one proach outperforms previous approaches on language to another preserves the geometric ar- the bilingual lexicon induction and cross- rangement of word embeddings. In a supervised lingual word similarity tasks. setting, the transformation matrix, W, is learned We next generalize our framework to rep- given a small bilingual dictionary and their resent multiple languages in a common latent corresponding monolingual embeddings.
    [Show full text]
  • L2F/INESC-ID at Semeval-2017 Tasks 1 and 2: Lexical and Semantic Features in Word and Textual Similarity
    L2F/INESC-ID at SemEval-2017 Tasks 1 and 2: Lexical and semantic features in word and textual similarity Pedro Fialho, Hugo Rodrigues, Lu´ısa Coheur and Paulo Quaresma Spoken Language Systems Lab (L2F), INESC-ID Rua Alves Redol 9 1000-029 Lisbon, Portugal [email protected] Abstract For word similarity, we test semantic equiva- lence functions based on WordNet (Miller, 1995) This paper describes our approach to the and Word Embeddings (Mikolov et al., 2013). Ex- SemEval-2017 “Semantic Textual Similar- periments are performed on test data provided ity” and “Multilingual Word Similarity” in the SemEval-2017 tasks, and yielded compet- tasks. In the former, we test our approach itive results, although outperformed by other ap- in both English and Spanish, and use a proaches in the official ranking. linguistically-rich set of features. These The document is organized as follows: in Sec- move from lexical to semantic features. In tion2 we briefly discuss some related work; in particular, we try to take advantage of the Sections3 and4, we describe our systems regard- recent Abstract Meaning Representation ing the “Semantic Textual Similarity” and “Mul- and SMATCH measure. Although with- tilingual Word Similarity” tasks, respectively. In out state of the art results, we introduce se- Section5 we present the main conclusions and mantic structures in textual similarity and point to future work. analyze their impact. Regarding word sim- ilarity, we target the English language and 2 Related work combine WordNet information with Word Embeddings. Without matching the best The general architecture of our STS system is systems, our approach proved to be simple similar to that of Brychc´ın and Svoboda(2016), and effective.
    [Show full text]
  • Deep Text Is an Approach to Text Analytics That Adds Depth and Intelligence to Our Ability to Utilize a Growing Mass of Unstructured Text
    DEEP “One of the most thorough walkthroughs of text analytics ever provided.” —Jeff Catlin, CEO, Lexalytics TEXT Deep text is an approach to text analytics that adds depth and intelligence to our ability to utilize a growing mass of unstructured text. In this book, author Tom Reamy explains what deep text is and surveys its many uses and benefits. DEEP Reamy describes applications and development best practices, discusses business issues including ROI, provides how-to advice and instruction, and offers Data to Big Big(ger) Text and Add Media, Social From Get RealValue Overload, to Conquer Information USING TEXT ANALYTICS guidance on selecting software and building a text analytics capability within an organization. Whether you need to harness a flood of social media content or turn a mountain of business information into an organized and useful asset, Deep Text will supply the TEXTUSING TEXT ANALYTICS insights and examples you’ll need to do it effectively. to Conquer Information Overload, Get Real Value From Social Media, “Sheds light on all facets of text analytics. Comprehensive, entertaining and Add Big(ger) Text to Big Data and enlightening.” —Fiona R. McNeill, Global Product Marketing Manager, SAS “Reamy takes the text analytics bull by the horns and gives it the time and exposure it deserves.” —Bryan Bell, VP of Global Marketing, Expert System REAMY “I highly recommend Deep Text as required reading for those whose work involves any form of unstructured content.” —Jim Wessely, President, Advanced Document Sciences $59.50 TOM REAMY DEEP TEXT Praise for Deep Text “Remarkably useful—a must-read for anyone trying to understand text analytics and how to apply it in the real world.” —Jeff Fried, CTO, BA Insight “A much-needed publication around the largely misunderstood field of text analytics … I highly recommend Deep Text as required reading for those whose work involves any form of unstructured content.” —Jim Wessely, President, Advanced Document Sciences “Sheds light on all facets of text analytics.
    [Show full text]
  • Tech Trends 2014 Inspiring Disruption Contents
    Tech Trends 2014 Inspiring Disruption Contents Introduction | 2 Disruptors CIO as venture capitalist | 6 Cognitive analytics | 18 Industrialized crowdsourcing | 30 Digital engagement | 42 Wearables | 54 Enablers Technical debt reversal | 66 Social activation | 78 Cloud orchestration | 88 In-memory revolution | 100 Real-time DevOps | 112 Exponentials | 124 Appendix | 136 Tech Trends 2014: Inspiring Disruption Introduction ELCOME to Deloitte’s fifth annual Technology Trends report. Each year, we study the ever- Wevolving technology landscape, focusing on disruptive trends that are transforming business, government, and society. Once again, we’ve selected 10 topics that have the opportunity to impact organizations across industries, geographies, and sizes over the next 18 to 24 months. The theme of this year’s report is Inspiring Disruption. In it, we discuss 10 trends that exemplify the unprecedented potential for emerging technologies to reshape how work gets done, how businesses grow, and how markets and industries evolve. These disruptive technologies challenge CIOs to anticipate their potential organizational impacts. And while today’s demands are by no means trivial, the trends we describe offer CIOs the opportunity to shape tomorrow—to inspire others, to create value, and to transform “business as usual.” The list of trends is developed using an ongoing process of primary and secondary research that involves: • Feedback from client executives on current and future priorities • Perspectives from industry and academic luminaries • Research by alliance partners, industry analysts, and competitor positioning • Crowdsourced ideas and examples from our global network of practitioners As in prior years, we’ve organized the trends into two categories. Disruptors are areas that can create sustainable positive disruption in IT capabilities, business operations, and sometimes even business models.
    [Show full text]
  • Natural Language Processing and Machine Learning As Practical
    Natural Language Processing and Machine Learning as Practical Toolsets for Archival Processing Tim Hutchinson University Archives and Special Collections, University of Saskatchewan, Saskatoon, Canada Records Management Journal, Special Issue: Technology and records management: disrupt or be disrupted? Volume 30, Issue 2, https://doi.org/10.1108/RMJ-09-2019-0055 Abstract Purpose – This study aims to provide an overview of recent efforts relating to natural language processing (NLP) and machine learning applied to archival processing, particularly appraisal and sensitivity reviews, and propose functional requirements and workflow considerations for transitioning from experimental to operational use of these tools. Design/methodology/approach – The paper has four main sections. 1) A short overview of the NLP and machine learning concepts referenced in the paper. 2) A review of the literature reporting on NLP and machine learning applied to archival processes. 3) An overview and commentary on key existing and developing tools that use NLP or machine learning techniques for archives. 4) This review and analysis will inform a discussion of functional requirements and workflow considerations for NLP and machine learning tools for archival processing. Findings – Applications for processing e-mail have received the most attention so far, although most initiatives have been experimental or project based. It now seems feasible to branch out to develop more generalized tools for born-digital, unstructured records. Effective NLP and machine learning tools for archival processing should be usable, interoperable, flexible, iterative and configurable. Originality/value – Most implementations of NLP for archives have been experimental or project based. The main exception that has moved into production is ePADD, which includes robust NLP features through its named entity recognition module.
    [Show full text]