This repository aims to provide further information and content relevant to the replication and extension of the following research: Investigating the Effectiveness of Clustering for Story Point Estimation by Vali Tawosi et al.
This work serves as my masters capstone research at Rochester Institute of Technology during the Spring 2023 semester.
To install dependencies simply run: pip install -r requirements.txt.
python generate_lda_model.py [PATH TO DATASET] [NUMBER OF TOPICS]
For example: python generate_lda_model.py .\tawosi_dataset\ 10
[PATH TO DATASET] is required. [NUMBER OF TOPICS] is optional but significantly reduces time to completion when provided. If it is not provided, the best t-value will be computed (very time intensive).
The best model produced is ./models/lda_2265.model which can be used to run SP estimations.
python run_sp_estimation.py [PATH TO DATASET] [PATH TO LDA MODEL] [PATH TO SAVE RESULTS TO] [Cluster Building Strategy:'MAE', 'MdAE', or 'sil'] [Algorithm Variant: 'LHC-SE' or 'LHC-TC-SE']
For example: python .\run_sp_estimation.py .\tawosi_dataset\ .\models\lda_2265.model .\results_new\ MAE LHC-SE
If [PATH TO SAVE RESULTS TO] is not already created then the directory will be created for you.
Advisor: Zhe Yu
Original Authors: