Sentiment Classification of News Text Data Using Intelligent Model

Does media sentiment affect stock prices? Evidence from China's STAR market

Objective

This paper explores the impact of media sentiment on stock prices on the Shanghai Stock Exchange Science and Technology Innovation Board (hereinafter the STAR market) from a behavioral finance perspective.

Methods

We collect Baidu News coverage of STAR-listed firms as the text, and measure text sentiment using a machine learning-based text analysis technique. We then empirically examine the impact of media sentiment on STAR market stock prices from two aspects: IPO pricing efficiency and IPO first-day stock performance.

Results

(1) Media sentiment has no significant impact on IPO pricing efficiency, thus suggesting that institutional investors participating in such offerings are generally not affected by media sentiment. (2) Optimistic media sentiment has a positive impact on IPO first-day returns, which indicates that individual investors are more easily influenced by media sentiment and therefore likely to abandon their rational judgment. (3) Media sentiment had a greater impact on IPO first-day returns during the COVID-19 pandemic than those before it, which suggests that individual investors are more influenced by media sentiment during pandemics.

Discussion

Our findings deepen the understanding of stock price formation on the STAR market, which provide a statistical basis for formulating policy directions and investment strategies.

Sentiment Thesaurus, Synset and Word2Vec Based Improvement in Bigram Model for Classifying Product Reviews

Classifying product reviews is one of the tasks in Natural Language Processing by which the sentiment of the reviewer towards a product can be identified. This identification is useful for the growth of the business by increasing the number of satisfied customers through product quality improvement. Bigram models are more popular in performing this classification since it considers the occurrence of two words consecutively in the reviews. In the existing works on bigram models, semantically similar words to the words present in bigrams are not considered. As the reviewers use different words with the same meaning to express their feeling, we proposed improved bigram models in which semantically similar words to the words in bigrams are also used for classifying the reviews. In the proposed models, sentiment polarity thesaurus is constructed by including sentiment words and their synonyms. The combinations of constructed thesaurus, Synset and Word2Vec are used for extracting synonyms for the words in the reviews. Performance of the proposed models is compared with the traditional bigram model and state-of-the-art methods. It is observed from the results that our models are able to achieve better performance than traditional model and recent methods.