Enhancement of stock market forecasting using an improved fundamental analysis-based approach

被引:53
作者
Chen, Yuh-Jen [1 ]
Chen, Yuh-Min [2 ]
Lu, Chang Lin [2 ]
机构
[1] Natl Kaohsiung First Univ Sci & Technol, Dept Accounting & Informat Syst, Kaohsiung, Taiwan
[2] Natl Cheng Kung Univ, Inst Mfg Informat & Syst, Tainan, Taiwan
关键词
Stock market forecasting; Fundamental analysis; Data envelopment analysis (DEA); Naive Bayesian classifier; Ensemble classifier; FEATURE-SELECTION; SVM; PREDICTION; ADABOOST;
D O I
10.1007/s00500-016-2028-y
中图分类号
TP18 [人工智能理论];
学科分类号
140502 [人工智能];
摘要
Stock investment is regarded as a high-risk financial activity, in which life savings can be destroyed when investors fail to consider factors in stock price variation or do not master professional knowledge and experiences related to investment. To enhance decision-making quality and profitability for investors, numerous different methods for forecasting stock market prices have appeared. However, these forecasts usually do not appear as expected because of uncertainties in the stock market. Therefore, how to effectively use stock information to help investors make stock investment decisions has become a primary issue in stock investment. Based on the bottom-up approach that considers financial conditions of listed companies, industrial environment, macroeconomics, and financial news respectively, an improved fundamental analysis-based approach for stock market forecasting is developed for selecting optimal stocks from the stock market and predicting their future price trends to provide a reference for investor decisions. This study involves the following tasks: (1) design an improved fundamental analysis-based approach to stock market forecasting; (2) develop techniques related to fundamental analysis-based stock market forecasting, and (3) demonstrate and evaluate the proposed fundamental analysis-based approach to stock market forecasting. The improved fundamental analysis-based approach to stock market forecasting involves techniques such as calculating the weight of financial indicators, evaluating and selecting individual stocks, selecting financial news features, determining stock trading signals based on financial news, and forecasting stock price trend.
引用
收藏
页码:3735 / 3757
页数:23
相关论文
共 32 条
[1]
[Anonymous], P 37 HAW INT C SYST
[2]
[Anonymous], 2010, P 4 INT AAAI C WEBL
[3]
[Anonymous], 2010, Journal of Machine Learning Research, DOI DOI 10.5555/1756006.1859899
[4]
Is all that talk just noise? The information content of Internet stock message boards [J].
Antweiler, W ;
Frank, MZ .
JOURNAL OF FINANCE, 2004, 59 (03) :1259-1294
[5]
Twitter mood predicts the stock market [J].
Bollen, Johan ;
Mao, Huina ;
Zeng, Xiaojun .
JOURNAL OF COMPUTATIONAL SCIENCE, 2011, 2 (01) :1-8
[6]
MEASURING EFFICIENCY OF DECISION-MAKING UNITS [J].
CHARNES, A ;
COOPER, WW ;
RHODES, E .
EUROPEAN JOURNAL OF OPERATIONAL RESEARCH, 1978, 2 (06) :429-444
[7]
Training data selection for improving discriminative training of acoustic models [J].
Chen, Berlin ;
Liu, Shih-Hung ;
Chu, Fang-Hui .
PATTERN RECOGNITION LETTERS, 2009, 30 (13) :1228-1235
[8]
Feature selection for text classification with Naive Bayes [J].
Chen, Jingnian ;
Huang, Houkuan ;
Tian, Shengfeng ;
Qu, Youli .
EXPERT SYSTEMS WITH APPLICATIONS, 2009, 36 (03) :5432-5435
[9]
Cheng Z., 2010, J CONVERGENCE INFORM, V5, P67
[10]
CORTES C, 1995, MACH LEARN, V20, P273, DOI 10.1023/A:1022627411411