Building Sentiment Lexicons Research Articles

Automatically Building Financial Sentiment Lexicons While Accounting for Negation

Financial investors make trades based on available information. Previous research has proved that microblogs are a useful source for supporting stock market decisions. However, the financial domain lacks specific sentiment lexicons that could be utilized to extract the sentiment from these microblogs. In this research, we investigate automatic approaches that can be used to build financial sentiment lexicons. We introduce weighted versions of the Pointwise Mutual Information approaches to build sentiment lexicons automatically. Furthermore, existing sentiment lexicons often neglect negation while building the sentiment lexicons. In this research, we also propose two methods (Negated Word and Flip Sentiment) to extend the sentiment building approaches to take into account negation when constructing a sentiment lexicon. We build the financial sentiment lexicons by leveraging 200,000 messages from StockTwits. We evaluate the constructed financial sentiment lexicons in two different sentiment classification tasks (unsupervised and supervised). In addition, the created financial sentiment lexicons are compared with each other and with other existing sentiment lexicons. The best performing financial sentiment lexicon is built by combining our Weighted Normalized Pointwise Mutual Information approach with the Negated Word approach. It outperforms all the other sentiment lexicons in the two sentiment classification tasks. In the unsupervised sentiment classification task, it has, on average, a balanced accuracy of 69.4%, and in the supervised setting, a balanced accuracy of 75.1%. Moreover, the various sentiment classification tasks confirm that the sentiment lexicons could be improved by taking into account negation while building the sentiment lexicons. The improvement could be made by using one of the proposed methods to incorporate negation in the sentiment lexicon construction process.

Sentiment lexicons and non-English languages: a survey

The ever-increasing number of Internet users and online services, such as Amazon, Twitter and Facebook has rapidly motivated people to not just transact using the Internet but to also voice their opinions about products, services, policies, etc. Sentiment analysis is a field of study to extract and analyze public views and opinions. However, current research within this field mainly focuses on building systems and resources using the English language. The primary objective of this study is to examine existing research in building sentiment lexicon systems and to classify the methods with respect to non-English datasets. Additionally, the study also reviewed the tools used to build sentiment lexicons for non-English languages, ranging from those using machine translation to graph-based methods. Shortcomings are highlighted with the approaches along with recommendations to improve the performance of each approach and areas for further study and research.

책 리뷰 말뭉치를 활용한 형용사, 부사 감성 사전 개발

초록·키워드 목차 오류제보하기 This paper presents a method of lexicon building for sentiment analysis using book review data. We show how to build sentiment lexicons automatically using book review data that includes readers" evaluation scores, basic positive and negative sentiment word lists using a semantic orientation measure, and sentiment word lists using Word2Vec. Specifically, we show how to automatically generate a useful lexicon of adjectives and adverbs for sentiment analysis. In summary, the process we describe is as follows: Basic word lists of positive and negative sentiments are created by relating word frequencies with reader"s book review scores. The expanded word lists are then built using the vector space representation of words and their cosine distance. This is made possible by assuming that words of similar sentiment and semantically co-related words will tend to appear together in the same context, and that they are represented proximately in the vector space. #감성분석(sentiment analysis) #웹 수집(web scraping) #리뷰 말뭉치(review corpus) #의미적 지향성(Semantic Orientation) #워드투벡(Word2Vec) 1. 머리말2. 관련 연구3. 연구 내용4. 요약 및 결론참고문헌

Word of mouth quality classification based on contextual sentiment lexicons

Word of mouth (WOM), also known as the passing of information from person to person or opinionated text, has become the main information resource for consumers when making purchase decisions. Whether WOM is a valuable reference source for consumers making a purchase is determined by the quality of the WOM. WOM quality classification is useful in filtering significant WOM documents from insignificant ones, and helps consumers to make their purchase decisions more efficiently. When a consumer has a negative experience, a lower rating score and negative text are generally provided and vice versa. Regardless of the sentimental polarity, high-quality WOM (i.e. with a very high or very low rating score) has a stronger influence on consumer behavior than low-quality WOM (i.e. with a medium rating score). We build three contextual lexicons to maintain the relationship between words and their associated sentimental categories. We then apply the technique of preference vector modeling and evaluate our proposed approach by four classifiers. According to the experiments for the internet movie database (IMDb) polarity data set and hotels.com data set, the proposed contextual lexicon-concept-quality (CLCQ) and contextual lexicon-quality (CLQ) models outperform the benchmarks, i.e. the static first-sense SentiWordNet and average-sense SentiWordNet models. These results demonstrate that the proposed models can be used as a viable approach for WOM quality classification. The novel aspects of this paper are three-fold. Firstly, we focus on WOM quality classification instead of traditional sentimental polarity classification. Secondly, we build sentiment lexicons from the contextual information, which are adaptable to domains. Thirdly, we integrate these contextual sentiment lexicons with preference vector modeling for WOM quality classification and achieve an outstanding improvement.

Sentiment Embeddings with Applications to Sentiment Analysis

We propose learning sentiment-specific word embeddings dubbed sentiment embeddings in this paper. Existing word embedding learning algorithms typically only use the contexts of words but ignore the sentiment of texts. It is problematic for sentiment analysis because the words with similar contexts but opposite sentiment polarity, such as good and bad , are mapped to neighboring word vectors. We address this issue by encoding sentiment information of texts (e.g., sentences and words) together with contexts of words in sentiment embeddings. By combining context and sentiment level evidences, the nearest neighbors in sentiment embedding space are semantically similar and it favors words with the same sentiment polarity. In order to learn sentiment embeddings effectively, we develop a number of neural networks with tailoring loss functions, and collect massive texts automatically with sentiment signals like emoticons as the training data. Sentiment embeddings can be naturally used as word features for a variety of sentiment analysis tasks without feature engineering. We apply sentiment embeddings to word-level sentiment analysis, sentence level sentiment classification, and building sentiment lexicons. Experimental results show that sentiment embeddings consistently outperform context-based embeddings on several benchmark datasets of these tasks. This work provides insights on the design of neural networks for learning task-specific word embeddings in other natural language processing tasks.

Building sentiment lexicons based on recommending services for the Polish language

Sentiment analysis has become a prominent area of research in computer science. It has numerous practical applications; e.g., evaluating customer satisfaction, identifying product promoters. Many methods employed in this task require language resources such as sentiment lexicons, which are unavailable for the Polish language. Such lexicons contain words annotated with their emotional polarization, but the manual creation of sentiment lexicons is very tedious. Therefore, this paper addresses this issue and describes a new method of building sentiment lexicons automatically based on recommending services. Next, the built lexicons were used in the task of sentiment classification.

A word-emoticon mutual reinforcement ranking model for building sentiment lexicon from massive collection of microblogs

Recently, more and more researchers have focused on the problem of analyzing people's sentiments and opinions in social media. The sentiment lexicon plays a crucial role in most sentiment analysis applications. However, the existing thesaurus based lexicon building methods suffer from the coverage problems when faced with the new words and new meanings in social media. On the other hand, the previous learning based methods usually need intensive expert efforts for annotating training datasets or designing extraction patterns. In this paper, we observe that the graphical emoticons are good natural sentiment labels for the corresponding microblog posts and a word-emoticon mutual reinforcement ranking model is proposed to learn the sentiment lexicon from the massive collection of microblog data. We integrate the emoticons and candidate sentiment words in the microblogs to construct a two-layer graph, on which a random walk is run for extracting the top ranked words as a sentiment lexicon. Extensive experiments were conducted on a benchmark dataset with various topics. The results validate the effectiveness of the proposed methods in building sentiment lexicon from microblog data.