# Common human diseases prediction using machine learning based on survey data

**Authors:** Jabir Al Nahian, Abu Kaisar Mohammad Masum, Sheikh Abujar, Md. Jueal Mia  
**arXiv:** [2209.10750](https://arxiv.org/abs/2209.10750)  
**Date:** 2022-09-22  
**Summary page:** https://www.summarizepaper.com/en/paper/2209.10750v1/

## Key points

- The medical field needs to shift towards preventative measures
- There is a need to predict diseases based on symptoms
- Researchers aimed to predict human diseases based on given symptoms
- They conducted a survey among people with different symptoms and used several classification algorithms to train their model
- The performance evaluation matrices were used to measure the model's accuracy and precision
- Unique from previous studies, this research successfully predicted human diseases based solely on their given symptoms with high accuracy rates ranging from 68.7% -88.5%
- Among the four classifiers tested: Random Forests; Support Vector Machines (SVM); Logistic Regression; K-Nearest Neighbors (KNN), it was found that part classifier outperformed others.
- This study provides insights into how technology can be used to predict diseases based on symptoms which could lead to better preventative measures and early detection.
- Further research is needed to improve the accuracy of disease prediction models and develop more robust disease detection systems.

## Summary

The medical field has long focused on disease detection and treatment, but there is a growing need to shift towards preventative measures. With the rise of various diseases such as COVID-19, flu, migraine, lung disease, heart disease, kidney disease, diabetes, stomach disease, gastric issues, bone disease and autism, it is imperative to find ways to predict these illnesses based on their symptoms. In this analysis published in the Bulletin of Electr Eng & Inf Vol. 11 No. 6 in December 2022, researchers aimed to predict human diseases based on given symptoms. They conducted a survey among people with different symptoms and used several classification algorithms to train their model. The performance evaluation matrices were used to measure the model's accuracy and precision. Comparative analysis was also done by looking at three papers related to multiple diseases prediction: Chang et al., Bang et al., and Kunjir et al.. Unique from previous studies is that this research successfully predicted human diseases based solely on their given symptoms with high accuracy rates ranging from 68.7% -88.5%. Among the four classifiers tested: Random Forests; Support Vector Machines (SVM); Logistic Regression; K-Nearest Neighbors (KNN), it was found that part classifier outperformed others. Overall this study provides insights into how technology can be used to predict diseases based on symptoms which could lead to better preventative measures and early detection. Further research is needed to improve the accuracy of disease prediction models and develop more robust disease detection systems.