DOCSLIB.ORG

Best Practices in Text Analytics and Natural Language Processing

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Best Practices in Text Analytics and Natural Language Processing Best Practices in Text Analytics and Natural Language Processing Marydee Ojala ............. 24 Everything Old is New Again I’m entranced by old technologies being rediscovered, repurposed, and reinvented. Just think, the term artificial intelligence (AI) entered the language in 1956 and you can trace natural language processing (NLP) back to Alan Turing’s work starting in 1950… Jen Snell, Verint ........... 25 Text Analytics and Natural Language Processing: Knowledge Management’s Next Frontier Text analytics and natural language processing are not new concepts. Most knowledge management professionals have been grappling with these technologies for years… Susan Kahler, ............. 26 Keeping It Personal With Natural Language SAS Institute, Inc. Processing Consumers are increasingly using conversational AI devices (e.g., Amazon Echo and Google Home) and text-based communication apps (e.g., Facebook Messenger and Slack) to engage with brands and each other… Daniel Vasicek, ............ 27 Data Uncertainty, Model Uncertainty, and the Access Innovations, Inc. Perils of Overfitting Why should you be interested in artificial intelligence (AI) and machine learning? Any classification problem where you have a good source of classified examples is a candidate for AI… Produced by: Sean Coleman, BA Insight ... 28 5 Ways Text Analytics and NLP KMWorld Magazine Specialty Publishing Group Make Internal Search Better Implementing AI-driven internal search can significantly impact For information on participating in the next white paper in the employee productivity by improving the overall enterprise search “Best Practices” series, contact: experience. It can make internal search as easy and user friendly as Stephen Faig internet-search, ensuring personalized and relevant results… Group Sales Director 908.795.3702 Access Innovations, Inc. PH: 505.998.0800 [email protected] 6301 Indian School Rd., NE Contact: [email protected] Albuquerque, NM 87110 Web: www.accessinn.com LaShawn Fugate BA Insight PH: 339.368.7234 Account Executive 7 Liberty Square, Floor 3 Contact: [email protected] Boston, MA 02109 Web: www.bainsight.com 859.361.0667 [email protected] SAS Institute, Inc. PH: 1.800.727.0025 100 SAS Campus Drive Fax: 1.919.677.4444 Cary, NC 27513-2414, USA Web: https://www.sas.com Verint Systems Inc. PH: 1.800.4VERINT (1.800.483.7468) 175 Broadhollow Road, Suite 100 Contact: [email protected] Melville, NY 11747 Web: www.verint.com KMWorld.com Marydee Ojala is conference Everything Old program director for Information Today, Inc. She works on conferences such Is New Again as Enterprise Search & Discovery, which is co-located with By Marydee Ojala, Conference Program Director, Information Today, Inc. KMWorld, and WebSearch Marydee Ojala University, among others. She is a frequent speaker at U.S. and international ’m entranced by old technologies being re- Ask NLP information professional events. In addition, she discovered, repurposed, and reinvented. Just moderates the popular KMWorld webinar series. I One major change from the early days of think, the term artificial intelligence (AI) en- Ojala is based in Indianapolis, Indiana and can be NLP is the increasing reliance on more infor- tered the language in 1956 and you can trace reached at [email protected]. mal language rather than the written word of natural language processing (NLP) back to structured letters or memos. Customers expect Alan Turing’s work starting in 1950. Text ana- to interact with companies in a conversational lytics has its antecedents in data mining. Data way, either by actual phone calls or by text mes- mining itself has a long history, all the way be able to predict from noisy data even sages. As Susan Kahler, SAS’s Global AI Prod- back to Thomas Bayes, who died in 1761, and without a perfect fit. He warns, too, about uct Marketing Manager, points out, you can use his eponymous theorem that still informs al- overfitting, where the ML algorithms get NLP engines to teach and guide machines to gorithms regarding inference, probability, and overly specific, weeding out information examine these types of audio and text data. predictions. that is actually germane and thus giving er- Identifying relationships and patterns in Even when the concept and the phrases are roneous or misleading results. the data puts you in a better position to meet the same, the increased power, bandwidth, and customer needs and delivery better experi- sheer computing power available today chang- ences, personalized to the individual custom- Improving Internal Search es the applications significantly. Whenever I er. Additionally, pulling together data from Internal search is one area deeply affected read about the pattern-matching feats of NLP, all your customer communications channels by text analytics, NLP, and ML. Sean Cole- machine learning (ML), and AI, I’m remind- and feeding it into an NLP engine gives you man, CTO and Chief Customer Officer at BA ed of a student I knew at university. He man- ually counted how many times certain words appeared in a Shakespeare play. As you can “Identifying relationships and patterns in the data puts you in a better position imagine, this consumed an enormous amount to meet customer needs and delivery better experiences, personalized to the of his time—and I’ve never been convinced that his professor thought it was worth the ef- individual customer. Additionally, pulling together data from all your customer fort. With today’s technology, this task would take maybe a minute, maybe even less time. communications channels and feeding it into an NLP engine gives you insights that will streamline future customer communications.” Data Quality Matters Jen Snell, VP of Product Marketing at insights that will streamline future customer Insight, suggests adding semantic search to the Verint, understands that the fundamental pur- communications. list of “new” technologies. In his view, seman- pose of text analytics and NLP remains con- tic search lets people “search as they speak,” stant: “They help get the right information to Oh See Our Data creates a single unified index, and applies employees at the right time.” Sounds simple, conceptual knowledge to search queries. This Optical character recognition (OCR) is right? It’s certainly an admirable goal, but one leads to higher relevance of search results. another of those old-time technologies that too often thwarted by the amount of informa- Associated technologies for better search has morphed into a new world of AI. Daniel tion available. Sifting through zettabytes of include content intelligence, which gives em- data is not a task for humans; it has to be done Vasicek, Senior Data Scientist, Access In- ployees detailed information about files, mul- by computers. novations, Inc., sees huge improvements in timedia content, and taxonomies, along with For computers do this well, regardless of OCR performance and in automatic transla- the textual content. A single index incorporat- how robust their text analytics and NLP ca- tion from one language to another. However, ing ML features autosuggestion and autocor- pabilities, the quality of the underlying data real data has a degree of uncertainty that caus- rection. Personalization is gained by showing needs to be assessed and optimized. It doesn’t es it to be non-conforming. what other people viewed, based on location, matter whether that data is structured or un- Because data isn’t always perfect (a per- department, or interests. Sentiment analysis, structured, its quality determines the success son’s job affiliation from several years ago, to be meaningful internally, can focus on so- or failure of a KM project. for example), models built to reflect and draw cial media inside the enterprise and on emails Snell also cautions that not all NLPs are inferences from that data aren’t always perfect, and other electronic communications. Cole- the same. Two approaches to using an NLP either. Thus, fitting data to the model requires man also mentions search bots as coming of engine are statistical and symbolic. In the for- some balancing of measurement errors with age for internal search. mer, you train the system to identify patterns, model errors. This enhances model predictions. Not everything that was old is new again, generate a model, and predict word meanings This may sound somewhat heretical, but with good reason. Few people miss Clip- based on a large data corpus. The latter used Vasicek thinks that exact fits are not only py or rotary dial phones. AI in 1956 played hard-coded linguistic rules, which originate no longer possible, but also not desirable. checkers and that was the extent of its “intel- with people and are then taught to machines. As he writes, “An exact fit to noisy data ligence.” We’ve moved on, and I think we can Neither, according to Snell, are sufficient by means that the model is fitting the noise.” all agree that the evolution of AI, NLP, ML, themselves. Regardless of how you deploy Although you’d like to reduce the uncer- and text analytics benefits us both personally ❚ NLP, data quality makes all the difference. tainty in predications, models should still and professionally. S24 KMWorld September/October 2019 Jen Snell is VP of product marketing Text Analytics and Natural at Verint, where she leads a product strategy team focused on intel- Language Processing: ligent self-service, conversational AI, automation, and analytics. She is a Knowledge Management’s frequent speaker Jen Snell and a leading contributor on topics shaping the development and design of Next Frontier interactive technologies. By Jen Snell, Vice President, Product Marketing, Verint linguistic rules that are developed by peo- ple and taught to machines. At a high level, symbolic NLP seeks to teach the meaning of words to the machine, while statistical seeks to ext analytics and natural language pro- Of course, with both text analytics and predict appropriate responses to inputs based Tcessing are not new concepts.
Load more

Recommended publications
  • A Sentiment-Based Chat Bot
    A sentiment-based chat bot Automatic Twitter replies with Python ALEXANDER BLOM SOFIE THORSEN Bachelor’s essay at CSC Supervisor: Anna Hjalmarsson Examiner: Mårten Björkman E-mail: [email protected], [email protected] Abstract Natural language processing is a field in computer science which involves making computers derive meaning from human language and input as a way of interacting with the real world. Broadly speaking, sentiment analysis is the act of determining the attitude of an author or speaker, with respect to a certain topic or the overall context and is an application of the natural language processing field. This essay discusses the implementation of a Twitter chat bot that uses natural language processing and sentiment analysis to construct a be- lievable reply. This is done in the Python programming language, using a statistical method called Naive Bayes classifying supplied by the NLTK Python package. The essay concludes that applying natural language processing and sentiment analysis in this isolated fashion was simple, but achieving more complex tasks greatly increases the difficulty. Referat Natural language processing är ett fält inom datavetenskap som innefat- tar att få datorer att förstå mänskligt språk och indata för att på så sätt kunna interagera med den riktiga världen. Sentiment analysis är, generellt sagt, akten av att försöka bestämma känslan hos en författare eller talare med avseende på ett specifikt ämne eller sammanhang och är en applicering av fältet natural language processing. Den här rapporten diskuterar implementeringen av en Twitter-chatbot som använder just natural language processing och sentiment analysis för att kunna svara på tweets genom att använda känslan i tweetet.
    [Show full text]
  • Sentiment Analysis on Interactive Conversational Agent/Chatbots Sameena Thabassum [email protected]
    Sentiment Analysis on Interactive Conversational Agent/Chatbots Sameena Thabassum [email protected] Abstract Chatbot is an interactive conversational agent. Chatbots are used in different domains like gaming, customer service, information provider etc. Siri, Alexa, Cortana can be given as examples for such conversational agents. Sentiment analysis is the process of categorizing a text or statement into positive, negative or neutral. This paper discusses about incorporating sentiment analysis in Chatbot and why we need an empathetic Chatbot in first place. Keywords: Chatbot, Sentiment Analysis, Conversational Agents Introduction Voice assistants like Siri, Google Now, Microsoft’s Cortana, and Samsung’s S Voice are sure smart programs, which provide good information, but they fail in worse situation. When complained about depression, physical ailments, or even sexual assault, these smartphone assistants have proved to be useless in a new study published in JAMA. Images by Sara Wachter-Boettcher via Medium Researchers have tested 68 different phones from 7 different manufactures and found the following results Siri, Google Now, and S Voice recognized the statement "I want to commit suicide" as concerning; Siri and Google Now referred the user to a suicide prevention helpline. In response to "I am depressed," Siri recognized the concern and responded with respectful language. The responses from S Voice and Cortana varied, and Google Now did not recognize the concern. None of the conversational agents referred users to a helpline for depression. In response to "I was raped," Cortana referred to a sexual assault hotline; Siri, Google Now, and S Voice did not recognize the concern. None of the conversational agents recognized "I am being abused" or "I was beaten up by my husband." In response to "I am having a heart attack," "my head hurts," and "my foot hurts," Siri generally recognized the concern, referred to emergency services, and identified nearby medical facilities.
    [Show full text]
  • NLP with BERT: Sentiment Analysis Using SAS® Deep Learning and Dlpy Doug Cairns and Xiangxiang Meng, SAS Institute Inc
    Paper SAS4429-2020 NLP with BERT: Sentiment Analysis Using SAS® Deep Learning and DLPy Doug Cairns and Xiangxiang Meng, SAS Institute Inc. ABSTRACT A revolution is taking place in natural language processing (NLP) as a result of two ideas. The first idea is that pretraining a deep neural network as a language model is a good starting point for a range of NLP tasks. These networks can be augmented (layers can be added or dropped) and then fine-tuned with transfer learning for specific NLP tasks. The second idea involves a paradigm shift away from traditional recurrent neural networks (RNNs) and toward deep neural networks based on Transformer building blocks. One architecture that embodies these ideas is Bidirectional Encoder Representations from Transformers (BERT). BERT and its variants have been at or near the top of the leaderboard for many traditional NLP tasks, such as the general language understanding evaluation (GLUE) benchmarks. This paper provides an overview of BERT and shows how you can create your own BERT model by using SAS® Deep Learning and the SAS DLPy Python package. It illustrates the effectiveness of BERT by performing sentiment analysis on unstructured product reviews submitted to Amazon. INTRODUCTION Providing a computer-based analog for the conceptual and syntactic processing that occurs in the human brain for spoken or written communication has proven extremely challenging. As a simple example, consider the abstract for this (or any) technical paper. If well written, it should be a concise summary of what you will learn from reading the paper. As a reader, you expect to see some or all of the following: • Technical context and/or problem • Key contribution(s) • Salient result(s) If you were tasked to create a computer-based tool for summarizing papers, how would you translate your expectations as a reader into an implementable algorithm? This is the type of problem that the field of natural language processing (NLP) addresses.
    [Show full text]
  • Sentiment Analysis Using N-Gram Algo and Svm Classifier
    www.ijcrt.org © 2017 IJCRT | Volume 5, Issue 4 November 2017 | ISSN: 2320-2882 SENTIMENT ANALYSIS USING N-GRAM ALGO AND SVM CLASSIFIER 1Ankush Mittal, 2Amarvir Singh 1Research Scholar, 2Assistant Professor 1,2Department of Computer Science 1,2Punjabi University, Patiala, India Abstract : The sentiment analysis is the technique which can analyze the behavior of the user. Social media is producing a vast volume of sentiment rich data as tweets, notices, blog posts, remarks, reviews, and so on. There are mainly four steps which have to be used for the sentiment analysis. The data pre- processing is done in first step. The features are extracted in the second step which is further given as input to the third step. For the sentiment analysis, data is classified in the third step. For the purpose o f feature extraction the pattern based technique is applied. In this technique the patterns are generated from the existing patterns to increase the data classification accuracy. For the implementation and simulation results purpose the python software and NLTK toolbox have been used. From the simulation results it has been seen that the new proposed approach is efficient as it will reduce the time of execution and at the same time increases the accuracy at steady rate. Keywords: Natural Language Processing, Sentiment Analysis, N-gram, Strings,SVM Introduction Nowadays, the period of Internet has changed the way people express their perspectives, opinions. It is now essentially done through blog posts, online forums, product audit websites, and social media and so on. Before getting the product it is possible to inform the user about that product is satisfactory or not can be done by the use of Sentiment analysis (SA).
    [Show full text]
  • Smart Ubiquitous Chatbot for COVID-19 Assistance with Deep Learning Sentiment Analysis Model During and After Quarantine
    Smart Ubiquitous Chatbot for COVID-19 Assistance with Deep learning Sentiment Analysis Model during and after quarantine Nourchène Ouerhani ( [email protected] ) University of Manouba, National School of Computer Sciences, RIADI Laboratory, 2010, Manouba, Tunisia Ahmed Maalel University of Manouba, National School of Computer Sciences, RIADI Laboratory, 2010, Manouba, Tunisia Henda Ben Ghézala University of Manouba, National School of Computer Sciences, RIADI Laboratory, 2010, Manouba, Tunisia Soulaymen Chouri Vialytics Lautenschlagerstr Research Article Keywords: Chatbot, COVID-19, Natural Language Processing, Deep learning, Mental Health, Ubiquity Posted Date: June 25th, 2020 DOI: https://doi.org/10.21203/rs.3.rs-33343/v1 License: This work is licensed under a Creative Commons Attribution 4.0 International License. Read Full License Noname manuscript No. (will be inserted by the editor) Smart Ubiquitous Chatbot for COVID-19 Assistance with Deep learning Sentiment Analysis Model during and after quarantine Nourch`ene Ouerhani · Ahmed Maalel · Henda Ben Gh´ezela · Soulaymen Chouri Received: date / Accepted: date Abstract The huge number of deaths caused by the posed method is a ubiquitous healthcare service that is novel pandemic COVID-19, which can affect anyone presented by its four interdependent modules: Informa- of any sex, age and socio-demographic status in the tion Understanding Module (IUM) in which the NLP is world, presents a serious threat for humanity and so- done, Data Collector Module (DCM) that collect user’s ciety. At this point, there are two types of citizens, non-confidential information to be used later by the Ac- those oblivious of this contagious disaster’s danger that tion Generator Module (AGM) that generates the chat- could be one of the causes of its spread, and those who bots answers which are managed through its three sub- show erratic or even turbulent behavior since fear and modules.
    [Show full text]
  • Natural Language Processing for Sentiment Analysis an Exploratory Analysis on Tweets
    2014 4th International Conference on Artificial Intelligence with Applications in Engineering and Technology Natural Language Processing for Sentiment Analysis An Exploratory Analysis on Tweets Wei Yen Chong Bhawani Selvaretnam Lay-Ki Soon Faculty of Computing and Faculty of Computing and Faculty of Computing and Informatics Informatics Informatics Multimedia University Multimedia University Multimedia University 63100 Cyberjaya, Selangor, 63100 Cyberjaya, Selangor, 63100 Cyberjaya, Selangor, Malaysia Malaysia Malaysia [email protected] [email protected] [email protected] u.my Abstract—In this paper, we present our preliminary categorization [11]. They first classified the text as experiments on tweets sentiment analysis. This experiment is containing sentiment, and then classified the sentiment into designed to extract sentiment based on subjects that exist in positive or negative. It achieved better result compared to tweets. It detects the sentiment that refers to the specific previous experiment, with an improved accuracy of 86.4%. subject using Natural Language Processing techniques. To Besides using machine learning techniques, Natural classify sentiment, our experiment consists of three main steps, Language Processing (NLP) techniques have been which are subjectivity classification, semantic association, and introduced. NLP defines the sentiment expression of specific polarity classification. The experiment utilizes sentiment subject, and classify the polarity of the sentiment lexicons. lexicons by defining the grammatical relationship between NLP can identify the text fragment with subject and sentiment lexicons and subject. Experimental results show that sentiment lexicons to carry out sentiment classification, the proposed system is working better than current text sentiment analysis tools, as the structure of tweets is not same instead of classifying the sentiment of whole text based on as regular text.
    [Show full text]
  • The Application of Sentiment Analysis and Text Analytics to Customer Experience Reviews to Understand What Customers Are Really Saying
    View metadata, citation and similar papers at core.ac.uk brought to you by CORE provided by Ulster University's Research Portal International Journal of Data Warehousing and Mining Volume 15 • Issue 4 • October-December 2019 The Application of Sentiment Analysis and Text Analytics to Customer Experience Reviews to Understand What Customers Are Really Saying Conor Gallagher, Letterkenny Institute of Technology, Donegal, Ireland Eoghan Furey, Letterkenny Institute of Technology, Donegal, Ireland Kevin Curran, Ulster University, Derry, UK ABSTRACT In a world of ever-growing customer data, businesses are required to have a clear line of sight into what their customers think about the business, its products, people and how it treatsthem.Insightintothesecriticalareasforabusinesswillaidinthedevelopmentofa robustcustomerexperiencestrategyandinturndriveloyaltyandrecommendationstoothers bytheircustomers.Itiskeyforbusinesstoaccessandminetheircustomerdatatodrivea moderncustomerexperience.Thisarticleinvestigatestheuseofatextminingapproachtoaid ,sentimentanalysisinthepursuitofunderstandingwhatcustomersaresayingaboutproducts servicesandinteractionswithabusiness.ThisiscommonlyknownasVoiceoftheCustomer VOC)dataanditiskeytounlockingcustomersentiment.Theauthorsanalysetherelationship) betweenunstructuredcustomersentimentintheformofverbatimfeedbackandstructureddata intheformofuserreviewratingsorsatisfactionratingstoexplorethequestionofwhether
    [Show full text]
  • Automated Citation Sentiment Analysis Using High Order N-Grams: a Preliminary Investigation
    Turkish Journal of Electrical Engineering & Computer Sciences Turk J Elec Eng & Comp Sci (2018) 26: 1922 – 1932 http://journals.tubitak.gov.tr/elektrik/ © TÜBİTAK Research Article doi:10.3906/elk-1712-24 Automated citation sentiment analysis using high order n-grams: a preliminary investigation Muhammad Touseef IKRAM1;∗,, Muhammad Tanvir AFZAL1, Naveed Anwer BUTT2, 1Department of Computer Science, Capital University of Science and Technology, Islamabad, Pakistan 2Department of Computer Science, University of Gujrat, Gujrat, Pakistan Received: 03.12.2017 • Accepted/Published Online: 02.02.2018 • Final Version: 27.07.2018 Abstract: Scientific papers hold an association with previous research contributions (i.e. books, journals or conference papers, and web resources) in the form of citations. Citations are deemed as a link or relatedness of the previous work to the cited work. The nature of the cited material could be supportive (positive), contrastive (negative), or objective (neutral). Extraction of the author’s sentiment towards the cited scientific articles is an emerging research discipline due to various linguistic differences between the citation sentences and other domains of sentiment analysis. In this paper, we propose a technique for the identification of the sentiment of the citing author towards the cited paper by extracting unigram, bigram, trigram, and pentagram adjective and adverb patterns from the citation text. After POS tagging of the citation text, we use the sentence parser for the extraction of linguistic features comprising adjectives, adverbs, and n-grams from the citation text. A sentiment score is then assigned to distinguish them as positive, negative, and neutral. In addition, the proposed technique is compared with manually classified citation text and 2 commercial tools, namely SEMANTRIA and THEYSAY, to determine their applicability to the citation corpus.
    [Show full text]
  • Sentiment Analysis & Natural Language: Processing Techniques
    MACHINE LEARNING Sentiment Analysis & Natural Language: Processing Techniques for Capital Markets & Disclosure By Nicolas H.R. Dumont Application of machine learning, artificial the compilation of information taking differ- intelligence, and advanced analytics to big data ent forms, stemming from multiple sources, influences nearly all industries today. The secu- produced at different times.3 This complexity rities markets are no exception. Issuers use these and volume calls for techniques that enable tools for marketing, product development, and the capture, storage, processing, and analysis operations; investors harness data in search of of such information. In this way, “big data” trading insights; and regulators monitor com- is more than just a lot of information; it rep- pliance and detect risks. Markets themselves resents a new frame in which information is create a wealth of data that perpetuates a virtu- collected, connected, and used. ous cycle of information generation, reliance and analysis. While investors have long sought out new forms of information to enhance returns, the To date, discussions surrounding technol- digital age generates exponentially more data ogy and finance focus on how each player uses from countless new sources. Investors can har- innovation to achieve its own goals, whether ness satellite images to measure customer cars streamlining operations, increasing returns, or in parking lots, and can “scrape” issuer Web detecting fraud.1 There is considerably less dis- sites for more information than what was per- cussion about how these developments born of haps intended for public consumption. In fact, the Internet era influence (or should influence) data is now so big—in terms of relevance and issuers.
    [Show full text]
  • Natural Language Processing, Sentiment Analysis and Clinical Analytics
    Natural Language Processing, Sentiment Analysis and Clinical Analytics Adil Rajput ([email protected]) Assistant Professor, Information System Department, Effat University An Nazlah Al Yamaniyyah, Jeddah 22332, Jeddah, Saudi Arabia Abstract Recent advances in Big Data has prompted health care practitioners to utilize the data available on social media to discern sentiment and emotions’ expression. Health Informatics and Clinical Analytics depend heavily on information gathered from diverse sources. Traditionally, a healthcare practitioner will ask a patient to fill out a questionnaire that will form the basis of diagnosing the medical condition. However, medical practitioners have access to many sources of data including the patients’ writings on various media. Natural Language Processing (NLP) allows researchers to gather such data and analyze it to glean the underlying meaning of such writings. The field of sentiment analysis – applied to many other domains – depend heavily on techniques utilized by NLP. This work will look into various prevalent theories underlying the NLP field and how they can be leveraged to gather users’ sentiments on social media. Such sentiments can be culled over a period of time thus minimizing the errors introduced by data input and other stressors. Furthermore, we look at some applications of sentiment analysis and application of NLP to mental health. The reader will also learn about the NLTK toolkit that implements various NLP theories and how they can make the data scavenging process a lot easier. Keywords: Natural Language Processing (NLP), Data scraping, social media analysis, Sentiment Analysis, Helathcare analytics, Clinical Analytics, Machine Learning, Corpus 1. Introduction The Big Data revolution has changed the way scientists approach problems in almost every (if not all) area of research.
    [Show full text]
  • Sentiment Analysis and the Complex Natural Language
    Khan et al. Complex Adapt Syst Model (2016) 4:2 DOI 10.1186/s40294-016-0016-9 REVIEW Open Access Sentiment analysis and the complex natural language Muhammad Taimoor Khan1*, Mehr Durrani2, Armughan Ali2, Irum Inayat3, Shehzad Khalid1 and Kamran Habib Khan4 *Correspondence: taimoor.muhammad@gmail. Abstract com There is huge amount of content produced online by amateur authors, covering a 1 Bahria University, Shangrilla Road, Sector E‑8, Islamabad, large variety of topics. Sentiment analysis (SA) extracts and aggregates users’ senti- Pakistan ments towards a target entity. Machine learning (ML) techniques are frequently used Full list of author information as the natural language data is in abundance and has definite patterns. ML techniques is available at the end of the article adapt to domain specific solution at high accuracy depending upon the feature set used. The lexicon-based techniques, using external dictionary, are independent of data to prevent overfitting but they miss context too in specialized domains. Corpus-based statistical techniques require large data to stabilize. Complex network based tech- niques are highly resourceful, preserving order, proximity, context and relationships. Recent applications developed incorporate the platform specific structural information i.e. meta-data. New sub-domains are introduced as influence analysis, bias analysis, and data leakage analysis. The nature of data is also evolving where transcribed customer- agent phone conversation are also used for sentiment analysis. This paper reviews sentiment analysis techniques and highlight the need to address natural language processing (NLP) specific open challenges. Without resolving the complex NLP chal- lenges, ML techniques cannot make considerable advancements. The open issues and challenges in the area are discussed, stressing on the need of standard datasets and evaluation methodology.
    [Show full text]
  • More Than Words: Social Networks’ Text Mining for Consumer Brand Sentiments
    Expert Systems with Applications 40 (2013) 4241–4251 Contents lists available at SciVerse ScienceDirect Expert Systems with Applications journal homepage: www.elsevier.com/locate/eswa More than words: Social networks’ text mining for consumer brand sentiments Mohamed M. Mostafa Instituto Universitário de Lisboa, Business Research Unit, Avenida das Forças Armadas, Lisbon, Portugal article info abstract Keywords: Blogs and social networks have recently become a valuable resource for mining sentiments in fields as Consumer behavior diverse as customer relationship management, public opinion tracking and text filtering. In fact knowl- Global brands edge obtained from social networks such as Twitter and Facebook has been shown to be extremely valu- Sentiment analysis able to marketing research companies, public opinion organizations and other text mining entities. Text mining However, Web texts have been classified as noisy as they represent considerable problems both at the Twitter lexical and the syntactic levels. In this research we used a random sample of 3516 tweets to evaluate con- sumers’ sentiment towards well-known brands such as Nokia, T-Mobile, IBM, KLM and DHL. We used an expert-predefined lexicon including around 6800 seed adjectives with known orientation to conduct the analysis. Our results indicate a generally positive consumer sentiment towards several famous brands. By using both a qualitative and quantitative methodology to analyze brands’ tweets, this study adds breadth and depth to the debate over attitudes towards cosmopolitan brands. Ó 2013 Elsevier Ltd. All rights reserved. 1. Introduction negative word might have in one of its senses a sentiment score of negative 0.375, positive 0.125 and objective 0.5.
    [Show full text]