- Angelina Amato
- Rebecca Breland
- Charlie Cryer
- Braedyn Edwards
- Magdeline Ng
Within the zip folder, you should find a datasets folder containing all of our different datasets that were used to train our different estimators. Since some of our estimators use standardized data and some do not, you will find subfolders in the datasets folder that contain those necessary datasets.
It is recommended that you download all the datasets at once; however, if you would like to download them one at a time, the datasets needed for each classifier will be listed below.
To see our data preprocessing, feature engineering, or would like to recreate our datasets from scratch, please refer to the Preprocessing folder. In this folder you will find two files: Data_cleaning.ipynb and Data_cleaning_reduced.ipynb. Upload these files to a Jupyter server with the datasets/other/cardio_train.csv dataset downloaded to run the file in full.
In the first file, you will find some of our initial data cleaning efforts that involved several diagrams and outlier removal.
The second file was our second attempt at data cleaning and preprocessing that does not contain many diagrams; however, it produces most of the datasets used in our classifiers as seen below.
For each classifier that we analyzed, you will find their individual instructions and respective code paths from the zip in this section.
- Location: CVD_code.zip
- Files:
- KNN PCA 95.ipynb
- Datasets:
- datasets/standardized/engineered_df_with_category_interactions.csv
To run KNN PCA 95.ipynb, upload the Jupyter Notebook file to a Jupyter server, as well as download the above mentioned dataset. You should then be able to run the file in its entirety.
- Location: CVD_code.zip
- Files:
- Logistic_Regression.ipynb
- Datasets:
- datasets/standardized/engineered_df_with_category_interactions.csv
- datasets/non standardized/cleaned_df.csv
To run Logistic_Regression.ipynb, upload the Jupyter Notebook file to a Jupyter server, as well as download the above mentioned datasets. You should then be able to run the file in its entirety.
- Location: CVD_code.zip/Decision Trees/
- Files:
- Decision Trees (CVD Prediction First Attempt).ipynb
- Decision Trees (CVD Prediction Second Attempt).ipynb
- Datasets:
- datasets/other/X.csv
- datasets/other/y.csv
- datasets/other/cardio_train.csv
To see the preliminary results of the Decision Tree on its first training, please use Decision Trees (CVD Prediction First Attempt).ipynb. Upload this Jupyter notebook to a Jupyter server and please have the cardio_train.csv dataset downloaded.
To see the final results of the Decision Tree, please use Decision Trees (CVD Prediction Second Attempt).ipynb, upload the Jupyter Notebook file to a Jupyter server, as well as upload the X.csv and y.csv files from above. You should then be able to run the file in its entirety.
- Location: CVD_code.zip/Random Forest/
- Files
- RandomForest_PCA.ipynb
- RandomForest_NoFeatEng.ipynb
- RandomForest_Demo.ipynb
- Datasets:
- datasets/non standardized/engineered_df_with_category_interactions.csv
- datasets/non standardized/cleaned_df.csv
To see the preliminary results of the Random Forest classifier without feature engineering, upload RandomForest_NoFeatEng.ipynb to a Jupyter server. Upload the cleaned_df.csv dataset to run the file in full.
To see the final results of the Random Forest classifier with feature engineering and some PCA testing, upload RandomForest_PCA.ipynb to a Jupyter server. Upload the engineered_df_with_category_interactions.csv dataset to run the file in full.
To see our demo that was presented to the class, upload RandomForest_Demo.ipynb to a Jupyter server. For this file to work, you will need to download both the above mentioned datasets.
- Location: CVD_code.zip
- Files:
- SVM_Heart_Prediction.ipynb
- SVM_project_hyperparam_tuning.ipynb
- Datasets:
- datasets/standardized/engineered_df_without_category_interactions.csv
- datasets/non standardized/cleaned_df.csv
To view and reproduce the initial results of the SVM classifier before hyper-parameter tuning, download SVM_Heart_Prediction.ipynb and upload it to a Jupyter server. Upload the engineered_df_without_category_interactions.csv dataset to run the file in full.
To see the hyper-parameter tuning and final results of the SVM classifier, download SVM_project_hyperparam_tuning.ipynb and upload it to a Jupyter server with the cleaned_df.csv to run the file in full.# CVD-Prediction