Now have the ability to save your loved ones for real, help them by diagnosing heart disease early on with the help of our Machine Learning Model!
Knowing someone who passed away due to heart complications, a team member brought up this issue during our ideation session. Upon doing some research, we realized the seriousness of heart failure and looked for any way to make a contribution. We found a dataset with factors that may affect mortality due to heart failure. We then tried to combine machine learning principles and came up with an idea to build a prediction model. This way, we could use ML to detect the possibility of heart failure in the early stages and help prevent the condition from worsening. If developed correctly, this technology could save millions of lives around the world.
Our project aims to predict mortality due to heart failure through Machine Learning Algorithms. This way, heart complications can be detected and treated at early stages.
We build the models using algorithm classifiers such as LogisticRegression, KNeighborsClassifier, DecisionTree Classifier, and Random Forest Classifier. We predicted an accuracy score, F-1 score, Receiver Operating Characteristic(ROC) score, and ROC_AUC score using each of these algorithm classifiers. We plotted a ROC curve for each of these models to conclude that Random Forest Classifier performs better than any other algorithm.
One of our biggest challenges was cleaning the dataset. Having limited knowledge about what is causing anomalies, we spent some quality time transforming our dataset to make it appropriate for predicting an accuracy score. Nonetheless, we were able to clean our dataset effectively and implement various algorithm classifiers to output a high accuracy score of the dataset.
As freshmen, we are proud that we were able to make an impactful submission in this small time frame. From the beginning, we just wanted to grow and learn, and even though we faced problems along the way, we managed to solve them as a team. We are also proud of the connections we made with peers and industry experts.
We came into this hackathon with no hacking experience and basic knowledge in Machine Learning. Along the way, we learned different libraries, functions and algorithms in ML. We also collaborated with each other in writing code and learned about the intricacies of working in a team.
In the future, we hope to perform more analysis on the data so that we are able to eliminate some factors that might not play a pivotal role in our prediction. This may help improve our accuracy. Moreover, we plan to further develop our website and present our project in an interactive way through our website.
We built Machine Learning Models using Python and Google Colab Notebook. We used Numpy, MatPlotLib, and Seaborn libraries. In addition, we also developed a website using HTML, CSS, and Javascript to expand our skills to front end development.