Heart attack is a common cause of death nowadays, not only old people but younger ones have also been victims of this. Although we have specialized doctors in this, yet the doctor to patient ratio is huge making our medical infrastructure inefficient. This is where Machine Learning comes in, our model will predict the chances of heart attack of a person based on various parameters. With an accuracy of 83.51 % with an precision of 84.31% .
On giving the required parameters our model predicts the possibility of heart attack. We have deployed our model on the internet.
We tried with multiple Machine Learning algorithms like KNN , Logistic Regression , Decision Tree and Random Forest Classifier and looking at the initial performance, we tried to improve the model performance by hyper tuning the parameter, feature selection of random forest classfier based model.
Random Forest Classifier Algorithm is a supervised algorithm classification method. In this algorithm, some trees form a forest. All trees in the Random Forest give the expected value of the class, and the class with the most votes is the model prediction. The three common methods are: Forest RI (random input selection). Forest RC (random mix). Combination of forest RI and forest RC. It can be used for both classification and regression problems, but good for dealing with classification problems and overcoming missing values.
Further, after selecting the best model we deployed it using flask and heroku.
The biggest challenge was selecting the most precise and accurate algorithm for our model.
We are quite proud that our model shows an accuracy of 83.51 % with an precision of 84.31% . And it is very easy to use as it is deployed globally on the internet.
We are planning to make it more interactive by integrating with various medical APIs to make the website more useful and to make it a complete website and not only a detector one. We also plan to create a mobile application too to make it more feasible for users to interact and keep track of their heart condition.
Use our webapp at - https://heart-attack-predictor-101.herokuapp.com/