Dear friends, you can find the code in the attachment, it makes use of the positive approximation to find reduct. The main function is the rough_fwred and it calls automatically the rough_lower to compute the positive approximation. Additionally, you can find the roughset code that computes the lower and upper approximation for a given data set. Note that this code only process numeric values and doesn't process string so you may use something like one-to-one transform to process string.
Hello Mr Hamed,thank you for the code I will test it as soon as possible, Thanks for sharing, I thougth that I am the only one interessted on rougt set but since many people are interessted too it is really kind from you to share the source code
If you need further help in rough set please don't hesitate to contact me, also you can read my paper in rough set titled "maximal limited similarity based rough set model"
I used your rough set functions (the main function rough_fwred which call sub function rough_lower) to extract just relevant features but the resulted matrix after applying rough_fwred reduce the accuracy of image classification. what I did is right and a any advice how to apply rough set to extract just relevant features to improve the percentage of image classification.