IRIS National Fair (My Project)

Disease Predictor - Machine Learning Project

Abstract :

The increasing spread of Internet technologies and mobile devices has created new opportunities for an online healthcare system. In certain cases, online medical aid or healthcare advice is easier and faster to understand than in-person assistance. People are frequently hesitant to visit a hospital or a doctor for mild ailments. In many situations, however, these mild symptoms might lead to serious health problems. Because internet health advice is readily available, it can provide consumers with a significant advantage. Additionally, present internet healthcare systems are unreliable and inaccurate. This system takes the user's symptoms as input and produces the illness as an output. The Naive Bayes Classifier is used to make predictions.

Problem statement :

The traditional diagnostic technique entails a patient seeing a doctor, undergoing several medical tests, and then reaching a decision. This procedure takes a long time. This project offers an automated illness prediction system that depends on user input to reduce the time necessary for the first step of diagnosing symptoms. The system takes the user's input and generates a list of possible illnesses.

Module description :

When the symptoms are supplied as input, the system will forecast the illness. The Naive Bayesian method will be used to forecast the illness. This approach, according to the literature review, achieves the highest accuracy for a bigger dataset. The collection includes illness classifications as well as symptoms for each condition. 70% of the dataset will be utilized as training data, while 30% will be used as test data. The dataset would be used for training and testing, and the required output would be achieved.

Naive Bayes Algorithm :

This algorithm takes the user's input and forecasts the most likely illness. The dataset and the machine learning method are used to do this. The Naive Bayesian method is used here, and it uses a probabilistic approach. I used Scikit to understand how to use the library and put it into practice. I utilized multinomial NB for this since there are numerous variations, i.e. multiple symptoms.

GUI :

For the user interface, I utilized the Tkinter package. Python's standard GUI library is Tkinter. When coupled with Tkinter, Python makes it simple to develop a graphical user interface. Tkinter provides a powerful object-oriented interface to the Tk GUI toolkit.

Project by Akshay Singh.