Big Data Prediction and Analysis of World Death Patterns

Human death prediction and analysis can help national departments cope with future trends, which suggests that different countries take appropriate measures to reduce their future deaths numbers. It is challenging to predict the future death trends based on existing data and determine the reasons behind the differences between death compositions in different countries. This paper uses linear regression to predict the world death patterns and uses SDI and K-means to analyze data to find other deep reasons. The main innovation is that we analyze a more comprehensive death pattern and include more death factors in machine learning model training compared with contemporary works. Firstly, we search the data from Kaggle and filter the data. In prediction part, a linear regression model based on world death dataset is trained. Then, we implement a MongoDB+ HTML structure to visualize the prediction results. In analysis part, the SDI and the elbow method in the K-means algorithm are used to study the underlying causes of more deaths in some countries. To verify the reliability of the prediction model, we use R2 value and interpretation error value to evaluate the model. The experimental results show that the death trends generated by the prediction model roughly correlate with the actual trends. Our analysis demonstrates that the economical background and geographical factors are crucial in influencing the death patterns.