A data mining approach to evaluate stock-picking strategies
暂无分享,去创建一个
A supervised learning model is introduced to evaluate the adequacy of a proposed parameter set in creating a stock-picking strategy. Since the stock market's behavior appears volatile, identifying rules that consistently predict stock movement can, at best, be challenging and perhaps unattainable. Hence, the relationship among the attributes used to predict stock behavior is often obscure, leading some speculators to falsely perceive associations among a set of disparate attributes when constructing stock-picking rules. A data mining-based model is used to evaluate the adequacy of proposed parameter set when constructing investment decision making rules. Thus, the rules are built by analyzing a combination of historical stock market data and related derived attributes. Different users can propose a separate set of parameters for constructing the decision rules representing distinct stock-picking strategies. The presented approach uses established principles from the Knowledge Discovery in Databases (KDD) process, data classification functions, and other data mining techniques.
[1] Paolo Giudici,et al. Applied Data Mining for Business and Industry , 2009 .
[2] B. Malkiel. The Efficient Market Hypothesis and Its Critics , 2003 .
[3] Alex Alves Freitas,et al. Comprehensible classification models: a position paper , 2014, SKDD.
[4] Gregory Piatetsky-Shapiro,et al. The KDD process for extracting useful knowledge from volumes of data , 1996, CACM.