Business - Text Analytics, Text Mining, and Sentiment Analysis

Business Intelligence and Analytics: Systems for Decision Support Text Analytics, Text Mining, and Sentiment Analysis 1) In the chapter's opening vignette, IBM's computer named Watson outperformed human game champions on the game show Jeopardy! Answer: TRUE 2) Text analytics is the subset of text mining that handles information retrieval and extraction, plus data mining. Answer: FALSE 3) In text mining, inputs to the process include unstructured data such as Word documents, PDF files, text excerpts, e-mail and XML files. Answer: TRUE 4) During information extraction, entity recognition (the recognition of names of people and organizations) takes place after relationship extraction. Answer: FALSE 5) Categorization and clustering of documents during text mining differ only in the preselection of categories. Answer: TRUE 6) Articles and auxiliary verbs are assigned little value in text mining and are usually filtered out. Answer: TRUE 7) In the patent analysis case study, text mining of thousands of patents held by the firm and its competitors helped improve competitive intelligence, but was of little use in identifying complementary products. Answer: FALSE 8) The bag-of-words model is appropriate for spam detection but not for text analytics. Answer: TRUE 9) Chinese, Japanese, and Thai have features that make them more difficult candidates for natural language processing. Answer: TRUE 10) Regional accents present challenges for natural language processing. Answer: TRUE 11) In the Hong Kong government case study, reporting time was the main benefit of using SAS Business Analytics to generate reports. Answer: TRUE 12) Detecting lies from text transcripts of conversations is a future goal of text mining as current systems achieve only 50% accuracy of detection. Answer: FALSE 13) In the financial services firm case study, text analysis for associate-customer interactions were completely automated and could detect whether they met the company's standards. Answer: TRUE 14) In text mining, creating the term-document matrix includes all the terms that are included in all documents, making for huge matrices only manageable on computers. Answer: FALSE 15) In text mining, if an association between two concepts has 7% support, it means that 7% of the documents had both concepts represented in the same document. Answer: TRUE 16) In sentiment analysis, sentiment suggests a transient, temporary opinion reflective of one's feelings. Answer: FALSE 17) Current use of sentiment analysis in voice of the customer applications allows companies to change their products or services in real time in response to customer sentiment. Answer: TRUE 18) In sentiment analysis, it is hard to classify some subjects such as news as good or bad, but easier to classify others, e.g., movie reviews, in the same way. Answer: TRUE 19) The linguistic approach to speech handles processes elements such as intensity, pitch and jitter from speech recorded on audio. Answer: FALSE 20) In the BBVA case study, text analytics was used to help the company defend and enhance its reputation in social media. Answer: TRUE 21) In the opening vignette, the architectural system that supported Watson used all the following elements EXCEPT A) massive parallelism to enable simultaneous consideration of multiple hypotheses. B) an underlying confidence subsystem that ranks and integrates answers. C) a core engine that could operate seamlessly in another domain without changes. D) integration of shallow and deep knowledge. Answer: C 22) According to a study by Merrill Lynch and Gartner, what percentage of all corporate data is captured and stored in some sort of unstructured form? A) 15% B) 75% C) 25% D) 85% Answer: D 23) Which of these applications will derive the LEAST benefit from text mining? A) patients' medical files B) patent description files C) sales transaction files D) customer comment files Answer: C 24) In text mining, stemming is the process of A) categorizing a block of text in a sentence. B) reducing multiple words to their base or root. C) transforming the term-by-document matrix to a manageable size. D) creating new branches or stems of recorded paragraphs. Answer: B 25) In text mining, tokenizing is the process of A) categorizing a block of text in a sentence. B) reducing multiple words to their base or root. C) transforming the term-by-document matrix to a manageable size. D) creating new branches or stems of recorded paragraphs. Answer: A 26) All of the following are challenges associated with natural language processing EXCEPT A) dividing up a text into individual words in English. B) understanding the context in which something is said. C) distinguishing between words that have more than one meaning. D) recognizing typographical or grammatical errors in texts. Answer: A 27) What application is MOST dependent on text analysis of transcribed sales call center notes and voice conversations with customers? A) finance B) OLAP C) CRM D) ERP Answer: C 28) In text mining, which of the following methods is NOT used to reduce the size of a sparse matrix? A) using a domain expert B) normalizing word frequencies C) using singular value decomposition D) eliminating rarely occurring terms Answer: B 29) What data discovery process, whereby objects are categorized into predetermined groups, is used in text mining? A) clustering B) association C) classification D) trend analysis Answer: C 30) In the research literature case study, the researchers analyzing academic papers extracted information from which source? A) the paper abstract B) the paper keywords C) the main body of the paper D) the paper references Answer: A 31) Sentiment classification usually covers all the following issues EXCEPT A) classes of sentiment (e.g., positive versus negative). B) range of polarity (e.g., star ratings for hotels and for restaurants). C) range in strength of opinion. D) biometric identification of the consumer expressing the sentiment. Answer: D 32) In sentiment analysis, which of the following is an implicit opinion? A) The hotel we stayed in was terrible. B) The customer service I got for my TV was laughable. C) The cruise we went on last summer was a disaster. D) Our new mayor is great for the city. Answer: B 33) In the Whirlpool case study, the company sought to better understand information coming from which source? A) customer transaction data B) delivery information C) customer e-mails D) goods moving through the internal supply chain Answer: C 34) What do voice of the market (VOM) applications of sentiment analysis do? A) They examine customer sentiment at the aggregate level. B) They examine employee sentiment in the organization. C) They examine the stock market for trends. D) They examine the "market of ideas" in politics. Answer: A 35) How is objectivity handled in sentiment analysis? A) It is ignored because it does not appear in customer sentiment. B) It is incorporated as a type of sentiment. C) It is clarified with the customer who expressed it. D) It is identified and removed as facts are not sentiment. Answer: D 36) Identifying the target of an expressed sentiment is difficult for all the following reasons EXCEPT A) the review may not be directly connected to the target through the topic name. B) blogs and articles with the sentiment may be general in nature. C) strong sentiments may be generated by a computer, not a person. D) sometimes there are multiple targets expressed in a sentiment. Answer: C 37) In text analysis, what is a lexicon? A) a catalog of words, their synonyms, and their meanings B) a catalog of customers, their words, and phrase C) a catalog of letters, words, phrases and sentences D) a catalog of customers, products, words, and phrase Answer: A 38) What types of documents are BEST suited to semantic labeling and aggregation to determine sentiment orientation? A) medium- to large-sized documents B) small- to medium-sized documents C) large-sized documents D) collections of documents Answer: B 39) Inputs to speech analytics include all of the following EXCEPT A) written transcripts of calls to service centers. B) recorded conversations of customer call-ins. C) live customer interactions with service representatives. D) videos of customer focus groups. Answer: A 40) In the Blue Cross Blue Shield case study, speech analytics were used to identify "confusion" calls by customers. What was true about these calls? A) They took less time than others as frustrated customers hung up. B) They led customers to rely more on self-serve options. C) They were not documented by customer service reps for speech analytics. D) They were difficult to identify using standard phrases like "I don't get it." Answer: C 41) IBM's Watson utilizes a massively parallel, text mining—focused, probabilistic evidencebased computational architecture called ________. Answer: DeepQA 42) ________ is probably the most often used form of information extraction. Answer: Named entity extraction 43) ________, also called homonyms, are syntactically identical words with different meanings. Answer: Polysemes 44) When a word has more than one meaning, selecting the meaning that makes the most sense can only be accomplished by taking into account the context within which the word is used. This concept is known as ________. Answer: word sense disambiguation 45) ________ is a technique used to detect favorable and unfavorable opinions toward specific products and services using large numbers of textual data sources. Answer: Sentiment analysis 46) In the text mining system developed by Ghani et al., treating products as sets of ________ rather than as atomic entities can potentially boost the effectiveness of many business applications. Answer: attribute-value pairs 47) In the Mining for Lies case study, a text based deception-detection method used by Fuller and others in 2008 was based on a process known as ________, which relies on elements of data and text mining techniques. Answer: message-feature mining 48) At a very high level, the text mining process can be broken down into three consecutive tasks, the first of which is to establish the ________. Answer: Corpus 49) Because the term-document matrix is often very large and rather sparse, an important optimization step is to reduce the ________ of the matrix. Answer: dimensionality 50) Where ________ appears in text, it comes in two flavors: explicit, where the subjective sentence directly expresses an opinion, and implicit, where the text implies an opinion. Answer: sentiment 51) ________ is mostly driven by sentiment analysis and is a key element of customer experience management initiatives, where the goal is to create an intimate relationship with the customer. Answer: Voice of the customer (VOC) 52) ________ focuses on listening to social media where anyone can post opinions that can damage or boost your reputation. Answer: Brand management 53) In sensitivity analysis, the task of differentiating between a fact and an opinion can also be characterized as calculation of ________ polarity. Answer: Objectivity-Subjectivity (OS) 54) When identifying the polarity of text, the most granular level for polarity identification is at the ________ level. Answer: word 55) When viewed as a binary feature, ________ classification is the binary classification task of labeling an opinionated document as expressing either an overall positive or an overall negative opinion. Answer: polarity 56) When labeling each term in the WordNet lexical database, the group of cognitive synonyms (or synset) to which this term belongs is classified using a set of ________, each of which is capable of deciding whether the synset is Positive, or Negative, or Objective. Answer: ternary classifiers 57) In automated sentiment analysis, two primary methods have been deployed to predict sentiment within audio: acoustic/phonetic and ________ modeling. Answer: linguistic 58) The time-demanding and laborious process of the ________ approach makes it impractical for use with live audio streams. Answer: acoustic/phonetic 59) ________ models operate on the premise that, when in a charged state, a speaker has a higher probability of using specific words, exclamations, or phrases in a particular order. Answer: Linguistic 60) Among the significant advantages associated with the ________ approach to linguistic modeling is the method's ability to maintain a high degree of accuracy no matter what the quality of the audio source, and its incorporation of conversational context through the use of structured queries. Answer: phonetic indexing and search 61) When IBM Research began looking for a major research challenge to rival the scientific and popular interest of Deep Blue, the computer chess-playing champion, what was the company's goal? Answer: The goal was to advance computer science by exploring new ways for computer technology to affect science, business, and society, and which would also have clear relevance to IBM business interests. 62) What is the definition of text analytics according to the experts in the field? Answer: Text analytics is a broader concept that includes information retrieval as well as information extraction, data mining, and Web mining. 63) How would you describe information extraction in text mining? Answer: Information extraction is the identification of key phrases and relationships within text by looking for predefined objects and sequences in text by way of pattern matching. 64) Natural language processing (NLP), a subfield of artificial intelligence and computational linguistics, is an important component of text mining. What is the definition of NLP? Answer: NLP is a discipline that studies the problem of "understanding" the natural human language, with the view of converting depictions of human language into more formal representations in the form of numeric and symbolic data that are easier for computer programs to manipulate. 65) In the security domain, one of the largest and most prominent text mining applications is the highly classified ECHELON surveillance system. What is ECHELON assumed to be capable of doing? Answer: Identifying the content of telephone calls, faxes, e-mails, and other types of data and intercepting information sent via satellites, public switched telephone networks, and microwave links 66) Describe the query-specific clustering method as it relates to clustering. Answer: This method employs a hierarchical clustering approach where the most relevant documents to the posed query appear in small tight clusters that are nested in larger clusters containing less similar documents, creating a spectrum of relevance levels among the documents. 67) Name and briefly describe four of the most popular commercial software tools used for text mining. Answer: • ClearForest offers text analysis and visualization tools. • IBM offers SPSS Modeler and data and text analytics toolkits. • Megaputer Text Analyst offers semantic analysis of free-form text, summarization, clustering, navigation, and natural language retrieval with search dynamic refocusing. • SAS Text Miner provides a rich suite of text processing and analysis tools. • KXEN Text Coder (KTC) offers a text analytics solution for automatically preparing and transforming unstructured text attributes into a structured representation for use in KXEN Analytic Framework. • The Statistica Text Mining engine provides easy-to-use text mining functionality with exceptional visualization capabilities. • VantagePoint provides a variety of interactive graphical views and analysis tools with powerful capabilities to discover knowledge from text databases. • The WordStat analysis module from Provalis Research analyzes textual information such as responses to open-ended questions, interviews, etc. • Clarabridge text mining software provides end-to-end solutions for customer experience professionals wishing to transform customer feedback for marketing, service, and product improvements. 68) Sentiment analysis has many names. Which other names is it often known by?Answer: Sentiment analysis is often referred to as opinion mining, subjectivity analysis, and appraisal extraction. 69) Identify, with a brief description, each of the four steps in the sentiment analysis process. Answer: 1. Sentiment Detection: Here the goal is to differentiate between a fact and an opinion, which may be viewed as classification of text as objective or subjective. 2. N-P Polarity Classification: Given an opinionated piece of text, the goal is to classify the opinion as falling under one of two opposing sentiment polarities, or locate its position on the continuum between these two polarities. 3. Target Identification: The goal of this step is to accurately identify the target of the expressed sentiment. 4. Collection and Aggregation: In this step all text data points in the document are aggregated and converted to a single sentiment measure for the whole document. 70) Within the context of speech analytics, what does the linguistic approach focus on? Answer: The linguistic approach focuses on the explicit indications of sentiment and context of the spoken content within the audio.