Data Mining: Applying and comparing the performance of classification algorithms on a clinical dataset
Date
2021
Authors
محمد, أبرار
عودة, دانيا
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Our project had 5 main achievements:
1. Applying different well-known prediction algorithms on a clinical dataset obtained
from Razan Center for Infertility which are:
Naïve Bayes Classifier
Random Forest
Logistic Regression
K-NN algorithm
2. Applying the RIMARC algorithm, which is a supervised (prediction) algorithm that
learns a scoring function to rank instances, developed and published as a
scientific research and an executable jar file.
3. Compare the performance of RIMARC and the other 4 prediction algorithm
regarding the same clinical data using 2 main indicators:
Area Under the Roc Curve indicator (AUC).
Execution time indicator.
4. Document the results of the previous mentioned experiment in a scientific
research under the title of: Applying Prediction Algorithms on ICSI Treatment Related
Data for Performance Comparison.
5. Develop a code from the logic of SERA, which is based on the ranking algorithm
RIMARC, to estimate the success rate for the couples going through ICSI
treatment to help both Razan Center for Infertility and patients with more
informed decisions on whether to go through with the treatment or not.