Skip to content

Latest commit

 

History

History
48 lines (32 loc) · 2.29 KB

README.md

File metadata and controls

48 lines (32 loc) · 2.29 KB
Raw

TEAM KTLM

This repository contains our solution to the BBDC 2024. Team KTLM consists of two members: Leonie Sieger and Louis Mozart KAMDEM.

Installation

To reproduce our solution, clone this repository to your local computer using:

git clone https://github.com/Louis-Mozart/BBDC.git && cd BBDC

Before running the code, ensure you have all the dependencies installed. You can install them using pip and the provided requirements.txt file.

pip install -r requirements.txt

How to Run?

To run our model and save the solution, ensure you have all the required data files (those provided by BBDC) in this directory, then execute:

python3 main.py --save_our_solution yes

This will then save our solution in the current directory with the name KTLM_solution.csv

If the data files are located in another directory, specify the paths using arguments --data1, --data2, and --prof_skeleton where data1 is the prof_data.csv file, data2 is the SessionData-all.csv file and prof_skeleton is the prof_skeleton.csv file.

python3 main.py --data1 path_to_prof_data.csv --data2 path_to_SessionData-all.csv --prof_skeleton path_to_prof_skeleton.csv --save_our_solution yes

Example

python3 main.py --data1  /home/dice/Desktop/BBDC/prof_data.csv --data2  /home/dice/Desktop/BBDC/SessionData-all.csv --prof_skeleton /home/dice/Desktop/BBDC/prof_skeleton.csv  --save_our_solution yes

To compute the score according to the BBDC solution, provide the prof_solution.csv file and execute:

python3 main.py --data1 prof_data.csv --data2 SessionData-all.csv --prof_skeleton prof_skeleton.csv --evaluate_score yes --prof_solution prof_solution.csv --save_our_solution yes

If the data are not in the same directory, replace prof_data.csv, SessionData-all.csv, prof_skeleton.csv and prof_solution.csv with their respective paths.

Hyperparameters

While we were unable to reproduce the same score achieved on the BBDC dashboard, we tuned our model to produce better and reproducible results. The hyperparameters we used for our model are printed during the preprocessing, training and postprocessing so please have a look on the terminal when the model is running.