Details

Autor(en) / Beteiligte

• Song, Chao
• Chen, Wanghu
• Fu, Lizhe
• Arshad, Ali

Titel

Improving Stock Price Prediction Based on Investing Sentiments

Ist Teil von

• 2020 IEEE 6th International Conference on Computer and Communications (ICCC), 2020, p.1639-1644

Ort / Verlag

IEEE

Erscheinungsjahr

2020

Quelle

IEEE

Beschreibungen/Notizen

• The change of stock prices has impacts on the sentiments of investors and vice versa. So, it will be a feasible way to improve the prediction of stock prices using investing sentiments. In this paper, stock prices are collected from a financial website and sentiments of investors to given stocks are also mining and evaluated from online comments. Two kinds of time series are constructed to reflect the fluctuation and trend going of the price of a stock respectively, as well as those of sentiments of investors to a specific stock. Taking the advantages of DCCA in time series analysis, Multiple Dimensional DCCA are proposed to evaluate the correlation between these two kinds of time series corresponding to stock prices and investing sentiments. Then, an approach to evaluate the impacting span of investing sentiments to the trend going of stock prices is proposed to enhance price prediction. Experiments show that when using the impacting span to the prediction of stock prices, the accuracy can be improved to about 85% stocks.