The web contains a wealth of product reviews, but sifting through them is a daunting task. Ideally, an opinion mining tool would process a set of search results for a given item, generating a list of product attributes (quality, features, etc.) and aggregating opinions about each of them (poor, mixed, good). We begin by identifying the unique properties of this problem and develop a method for automatically distinguishing between positive and negative reviews. Our classifier draws on information retrieval techniques for feature extraction and scoring, and the results for various metrics and heuristics vary depending on the testing situation. The best methods work as well as or better than traditional machine learning. When operating on individual sentences collected from web searches, performance is limited due to noise and ambiguity. But in the context of a complete web-based tool and aided by a simple method for grouping sentences into attributes, the results are qualitatively quite useful.

Sentiment classification is a recent subdiscipline of text classification which is concerned not with the topic a document is about, but with the opinion it expresses. It has a rich set of applications, ranging from tracking users ’ opinions about products or about political candidates as expressed in online forums, to customer relationship management. Functional to the extraction of opinions from text is the determination of the orientation of “subjective ” terms contained in text, i.e. the determination of whether a term that carries opinionated content has a positive or a negative connotation. In this paper we present a new method for determining the orientation of subjective terms. The method is based on the quantitative analysis of the glosses of such terms, i.e. the definitions that these terms are given in on-line dictionaries,

Many approaches to automatic sentiment analysis begin with a large lexicon of words marked with their prior polarity (also called semantic orientation). However, the contextual polarity of the phrase in which a particular instance of a word appears may be quite different from the word’s prior polarity. Positive words are used in phrases expressing negative sentiments, or vice versa. Also, quite often words that are positive or negative out of context are neutral in context, meaning they are not even being used to express a sentiment. The goal of this work is to automatically distinguish between prior and contextual polarity, with a focus on understanding which features are important for this task. Because an important aspect of the problem is identifying when polar terms are being used in neutral contexts, features for distinguishing between neutral and polar instances are evaluated, as well as features for distinguishing between positive and negative contextual polarity. The evaluation includes assessing the performance of features across multiple machine learning algorithms. For all learning algorithms except one, the combination of all features together gives the best performance. Another facet of the evaluation considers how the presence of neutral instances affects the performance of features for distinguishing between positive and negative polarity. These experiments show that the presence of neutral instances greatly degrades the performance of these features, and that perhaps the best way to improve performance across all polarity classes is to improve the system’s ability to identify when an instance is neutral. 1.

Ranking documents or sentences according to both topic and sentiment relevance should serve a critical function in helping users when topics and sentiment polarities of the targeted text are not explicitly given, as is often the case on the web. In this paper, we propose several sentiment information retrieval models in the framework of probabilistic language models, assuming that a user both inputs query terms expressing a certain topic and also specifies a sentiment polarity of interest in some manner. We combine sentiment relevance models and topic relevance models with model parameters estimated from training data, considering the topic dependence of the sentiment. Our experiments prove that our models are effective. 1

Information is an essential component of all financial markets and transactions. However information can arrive in multiple forms. In this paper, I begin to define what is meant by hard and soft information. Hard information is quantitative, easy to store and transmit in impersonal ways, and its content is independent of the collection process. Technology is changing the way we communicate and thus must fundamentally change the way financial markets and institutions operate. One of these changes is a greater reliance on hard relative to soft information in financial transactions. This paper discusses the advantages and costs of this substitution and the possible consequences of the hardening of information on both financial markets and institutions as well as those who study them.

In many applications, there is often a need for comparing multiple text collections to find commonalities and differences in topical themes, a task we refer to as comparative text mining. In this paper, we present a general comparative mining system (CTMS). The CTMS system takes any two collections of text and generates a list of cross-collection themes and their associated individual collection-specific themes. The themes are linked to representative passages in each collection. The themes are represented as word distributions, and the underlying comparative mining algorithm is based on a probabilistic mixture model. The system carries out all the stages of text mining from data cleaning and preprocessing to the actual mining and post-processing, allowing users to perform comparative analysis between any two collections and navigate through the extracted theme space. This system can potentially be applied to a broad range of areas including opinion summarization, business intelligence, and summarization of text.

The business media play an active role in influencing stock prices. Statistically significant excess returns at the time of the publication of stock recommendations have been documented many times. Frequently these abnormal gains begin to accumulate long before the publication date. In most cases they reach their highs on the day the recommendations are disseminated to the public. With few exceptions a price reversal sets in shortly thereafter: Excess returns in recommended stocks are at least partially given up.

Extracting sentiments from unstructured text has emerged as an important problem in many disciplines. An accurate method would enable us, for example, to mine online opinions from the Internet and learn customers ’ preferences for economic or marketing research, or for leveraging a strategic advantage. In this paper, we propose a two-stage Bayesian algorithm that is able to capture the dependencies among words, and, at the same time, finds a vocabulary that is efficient for the purpose of extracting sentiments. Experimental results on the Movie Reviews data set show that our algorithm is able to select a parsimonious feature set with substantially fewer predictor variables than in the full data set and leads to better predictions about sentiment orientations than several state-of-the-art machine learning methods. Our findings suggest that sentiments are captured by conditional dependence relations among words, rather than

The business media play an active role in influencing stock prices. Statistically significant excess returns at the time of the publication of stock recommendations have been documented many times. Frequently these abnormal gains begin to accumulate long before the publication date. In most cases they reach their highs on the day the recommendations are disseminated to the public. With few exceptions a price reversal sets in shortly thereafter: Excess returns in recommended stocks are at least partially given up. Many stocks now enter a period of underperformance, earning significant negative returns. The return reversions indicate that such stock price reactions are due to price pressure from "naive " investors hoping to profit from the experts. However, most media lack any real information that is not yet reflected in stock prices. In short: There is no evidence that stock recommendations published in the media offer any systematic opportunity to outperform the market. The evidence leads to the opposite conclusion: That investors who follow such advice will lose in the long run.

An experimental automated dialogue system that plays the role of a crisis hotline dispatcher is currently developed. Besides controlling the communication flow, this system is able to retrieve information about crisis situations from user’s input. It offers a natural user interaction by the ability to perceive and respond to human emotions. The system has an emotion recognizer that is able to recognize the emotional loading from user’s linguistic content. The recognizer uses a database that contains selected keywords on a 2D “arousal ” and “valence ” scale. The output of the system provides not only the information about the user’s emotional state but also an indication of the urgency of his/her information regarding to crisis. The dialogue system is able to start a user friendly dialogue, taking care of the content, context and emotional loading of user’s utterances.