Abstract:
Microarray data are becoming a more essential source of gene expression data for interpretation and analysis. To improve the detection accuracy of tumors, the researchers try to use the lowest feasible collection of the most gene expression studies, and relevant gene expression patterns are found. The purpose of this article is to use a data mining strategy and an optimized feature selection method focused on a limited dense tree forest classifier to evaluate and forecast colon cancer data. More specifically, merging the “gain information” and “Grey wolf optimization” was incorporated as a feature selection approach into the random forest classifier, to improve the prediction model’s accuracy. Our suggested technique can decrease the load of high-dimensional data, and it allows quicker computations. In this research, we provided a comparison of the analysis model with feature selection accuracy over model analysis without feature selection accuracy. The extensive experimental findings have shown that the suggested method with selecting features is beneficial, outperforming the good classification performances.
