Intelligent Feature Selection Research Articles

Multivariate Data Driven Models for Firm Value Prediction from Small Data Samples

Firm value is an essential outcome of firm managerial performance measure. From shareholders' and potential investors' perspectives it is an investment decision criterion. This paper studies how firm values are determined by our previous work that explored the variables that drive firm value growth in the Miscellaneous Industrials sector of the Australian Stock Market. We previously studied how the firm's value growth as well as its growth in relation to the changes in the Miscellaneous Industrial Index changed using traditional and artificial intelligent (AI) feature selection techniques. Previously multi-domain models were used to reduce the number of attributes to a manageable number. The primary aim of this study is to investigate the effect of macroeconomic and firm- specific economic factors on firm value growth and relative firm value growth models (also known as multi-domain models) performance. We also explore the capabilities of the multi-domain models to make out-of-sample predictions for firm value growth and relative firm value growth. We find that the macroeconomic and firm specific factors can improve the performance of these multi-domain models, where their raw values generally have a more positive impact on the model performance. It was also found that the price to earning ratio was a consistent positive performer.

An empirical study of AI-based multiple domain models for selecting attributes that drive company wealth

This paper explores what attributes drive company wealth creation in the Miscellaneous Industrials sector of the Australian Stock Market. We examine traditional and artificial intelligent (AI) feature selection techniques, to select attributes that drive company wealth and observe if a multiple domain model outperforms single domain models with regards to predicting company wealth. Using a large number of calculated attributes, our empirical findings suggest that a multiple domain model was most effective. We find that Market Capitalisation, Debt Asset Ratio, ROI and Beta Value attributes are the main contributors to following the directional change in the shareholder value. This was using a J4.8 and GRNN multiple domain model.

Toward integrating feature selection algorithms for classification and clustering

This paper introduces concepts and algorithms of feature selection, surveys existing feature selection algorithms for classification and clustering, groups and compares different algorithms with a categorizing framework based on search strategies, evaluation criteria, and data mining tasks, reveals unattempted combinations, and provides guidelines in selecting feature selection algorithms. With the categorizing framework, we continue our efforts toward-building an integrated system for intelligent feature selection. A unifying platform is proposed as an intermediate step. An illustrative example is presented to show how existing feature selection algorithms can be integrated into a meta algorithm that can take advantage of individual algorithms. An added advantage of doing so is to help a user employ a suitable algorithm without knowing details of each algorithm. Some real-world applications are included to demonstrate the use of feature selection in data mining. We conclude this work by identifying trends and challenges of feature selection research and development.