Title :
Online mining in unstructured financial information: An empirical study in bulletin news
Author :
Chao Ma ; Xun Liang
Author_Institution :
Sch. of Inf., Renmin Univ. of China, Beijing, China
Abstract :
The Internet produces massive financial unstructured textual information every day. How to utilize these unstructured data effectively is a challenging topic. In the background of A share T+0 and stock option promoting in the China security market, we present a model to recognize the risk and investment opportunity according to the massive online financial textual information. Since the key word vector is in extremely high dimension space and critical in influencing the performance of our forecast models, a manifold learning method is firstly applied to reduce its dimension while keeping essential features. By utilizing financial event study, we secondly apply support vector machines to predict the news type and sentiment value. The model can achieve the intelligent and instant match between textual news and the reactions of stock market. Our results provide prompt supports for financial practitioners to make investment decisions no matter they are in long or in short positions in the market.
Keywords :
Internet; data mining; learning (artificial intelligence); stock markets; support vector machines; China security market; Internet; investment decisions; investment opportunity; manifold learning method; massive financial unstructured textual information; online mining; stock market; support vector machines; Companies; Data mining; Investment; Kernel; Manifolds; Predictive models; Support vector machines; decision support; event study; investment; manifold learning; online risk identification; support vector machine; unstructured financial information;
Conference_Titel :
Service Systems and Service Management (ICSSSM), 2015 12th International Conference on
Conference_Location :
Guangzhou
Print_ISBN :
978-1-4799-8327-8
DOI :
10.1109/ICSSSM.2015.7170151
