An Industrial Internet of Things Feature Selection Method Based on Potential Entropy Evaluation Criteria

In recent years, with the rapid development of industrial Internet of Things, the rapid growth of data has become a severe challenge and precious opportunity faced by many industries. The information society has entered the era of big data. Feature selection is frequently used to reduce the number of features in many applications of Internet of things, where data of high dimensionality are involved. To the best of our knowledge, a fewer researchers focus on the physical distribution of data and the anisotropy of the data characteristics. To this end, this paper introduces a novel feature selection approach based on potential entropy evaluation criteria (FMPE). The FMPE method considers the distribution of the data itself when measuring the importance of the feature. The data is mapped into a high-dimensional space which has better divisibility by extending data field to generalized multidimensional data field. Related experiments and analyses on UCI data sets and face data sets show that the FMPE algorithm can effectively eliminate the unimportant features or noise features to improve the performance of the classification algorithm. A high classification accuracy is achieved by the combination of the selected feature subset and a variety of classifiers and the FMPE algorithm is independent of the specific classifier.