reduce number of factors based on statistical method based on significance

I am looking for an expert to help reduce the number of factors in a dataset based on statistical methods. I would like to use the Chi-squared test, ANOVA, and/or T-test with a significance level of 0.05. The goal is to use the reduced number of factors to enhance predictive accuracy and simplify data interpretation within a fuzzy comparison matrix. Successful completion of this project should result in a dataset made of fewer factors that can be easily taken into account in a reasonable amount of time.