I want to discretize a continuous attribute, I applied a lot of methods but it's clear that i'm losing information at each time, until I found the FDIC method wich seems great but I couldn't impelement it. This is the paper's name :Dynamic Data Discretization Technique based on Frequency and K-Nearest Neighbour algorithm . This is the explanation part from the paper: It begins with sorting each numerical attribute according to its value. The following is performed: 1. Compute frequency distribution of pattern and put into its own k interval; 2.Calculate the w intervals width (length) on each pattern observed by finding the distance between two k interval patterns; 3. Compute means of w intervals on the distribution of k intervals; 4. Automatic min threshold is computed based on means distribution


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.