A data challenge competition held by International Federation of Classification Societies(IFCS). Aiming to find a (semi)automatic classification of the lower back pain patients based on the baseline variables in order to find clinically applicable and useful groups.
My teammate and I are submitting our cluster analysis contribution in response to the data challenge. You'll find the following files:
- "LBP_Project_Report.pdf" — this is our written report which describes the entire process and methods used in our analysis.
- "Tokyo_presentation.pdf" - this is the presentation ppt that we are given at IFCS 2017 conference.
- "consolFullOrdinal.Rdata" — this is our imputed dataset which we use to perform cluster analysis.
- "LBP_ProjectCode.R" — this is our code for the data imputation, clustering, and other analysis. Running this code will generate 4 csv files saved to current working directory (two files for variable contributions, one file for the selected important variables table, and one file for cluster membership based on the clustering result).
- "Alg_Comparison.R" — we used this file to compare different clustering algorithms only and don’t need to run it since the final selected algorithm is already used in the "LBP_ProjectCode.R".
The Excel files "variables_data_challenge.xlsx" and "data_challenge.xlsx" are the raw data that was given by IFCS. Please put them in the current working directory in R, along with “consolFullOrdinal.Rdata” file, when the project code file is run.