This project represents my second independent project. It focuses on predicting fetal health based on a well feature engineered dataset containing essential details. The goal was to classify fetal health into three categories: 'normal', 'suspect', and 'pathological'. Given the nature of the data, it was clear that this problem could be approached as a classification task.
Choice of Algorithm: After careful consideration, a decision tree classifier was chosen for its simplicity and interpretability. Implementing this classifier proved to be straightforward and effective, especially given the dataset's structured nature.
Focus on Recall: Medical predictions demand a higher level of accuracy, especially in identifying potential health issues. To address this, the model's validation was focused primarily on the recall metric. This emphasis ensures that the model is particularly adept at capturing instances of fetal health issues, enhancing its reliability in a clinical context.
- numpy
- pandas
- sklearn
- matplotllib
- seaborn
- pickle
pip install numpy
pip install pandas
pip install sklearn
pip install matplotlib
pip install seaborn
To download the 'Fetal health' dataset 👉🏽 click here
- I seek improvement in data visualization and prediction capabilities.