techniques for in-depth mining (Miao 2018). Tian
investigated the association between investor
sentiment and stock prices in 2019 (Faxiang 2019).
Fan explored the heterogeneity of the relationship
between investor mood and stock returns of
individual stocks in different market capitalization
sizes, industries, and market states in 2021 (Pengying
2021). Yang demonstrated in 2022 that the investor
sentiment index is better for projecting stocks of
small and medium-sized, difficult-to-value
companies in the manufacturing industry (Xiaoyu
2022).
It has been noted that the news media adopts
shocking terms that may hurt investors. Sentiments
generated by social media such as news and stock bar
forums can spread rapidly and expand in the market,
thus significantly affecting the price formation
mechanism (Tetlock et al. 2008). In addition, these
sentiments play an important role in the formation of
investors' subjective beliefs and have a subtle
influence on their perceptions, judgments, and
decisions about stock investments (Zhu et al. 2017).
Eastmoney has a leading position in the field of
stock trading, its stock bar section has a rich exchange
of investor insights, the daily exchange of a huge
amount of data, which contains a large number of
stockholders on the stock's first emotional reaction,
by analyzing their emotions, the paper can achieve
part of the stock prediction.
In this paper, we construct the sentiment factor
through the methods of text analysis and sentiment
analysis, and through the study of the sentiment
factor, we observe its effect on the excess return of
the stock market. Finally, a regression model is used
to compare the prediction effect with and without the
sentiment factor, and it is concluded that the
prediction model is more stable and accurate when it
contains the sentiment factor. The research objective
of this study is to provide investors with a more
effective basis for decision-making.
2 DATA COLLECTION
2.1 Sample Period Selection
This paper chooses 2022 as the sample, the sample
selected from December 31, 2021, to December 31,
2022 trading data. Each stock may have multiple up
and down fluctuations during the one-year period,
which is reflected in the more obvious stockholder
sentiment, and there is a longer period, which is
conducive to analysis through market sentiment and
has better support for the comprehensiveness of the
sample.
2.2 Sample Stock Selection
This paper investigates the top 25 popular stocks in
Oriental Finance Network, Oriental Finance Network
is the current domestic daily browsing volume at the
forefront of the financial website, its popular stock list
can reflect the market attention to a certain extent,
such as the stock is valued, its liquidity is higher, the
market trading is active. It is easy to be influenced by
market sentiment, especially in the period of
following up and down, the corresponding heat of
discussion will be reflected more obviously, so
choose the stocks in the hot list for research.
First of all, this paper uses the requests library in
Python to crawl the popular page of the Oriental
finance network to obtain its top 25 popular stocks
and the corresponding code, and then through the
Pandas. data frame function to save it to Excel to
build the corresponding stock selection form. As
shown in Table 1, some representative stocks are
listed.
Table 1: Selected stocks.
Name
Code Name Code Name Code
Yawei Shares 002559 Haima Moto