Prediction of Angiographic Disease Status using Rule Based Data Mining Techniques

Author: Shabia Shabir Khan and S.M.K. Quadri

Download PDF

Abstract

Data mining is the process of uncovering the fluctuating hidden patterns or trends in the data that is not immediately apparent by just summarizing the data. It can help in predicting the future (predictive analytics) in addition to explain the current or past situation (descriptive analytics). After the interpretation of information, knowledge can be extracted by identifying relationships among patterns. Various data mining (machine learning) algorithms have been provided for extracting the nuggets of knowledge from medical datasets in the field of diagnostics. This paper discusses various machine learning techniques that have been evaluated using heart disease dataset for the prediction of class i.e. angiographic disease status (diameter narrowing). The main aim is to search a model that accurately predicts the class of the unknown records. The evaluation has been performed using WEKA software tool that helps in comparing the various techniques on the basis of certain important evalu

Keywords

Data mining, Machine Learning, Angiographic disease status, classification.

Conclusion

An overview has been presented to summarize the various data mining techniques that can help in efficient prediction for early medical diagnosis. This paper has experimentally proved that, for the same dataset, different algorithms work in different ways. As far as accuracy is concerned, the comparison between various rule based classifiers concluded neural network as an optimal model for classification in complex heart disease dataset. This is evident from the various evaluation measures like correctly or incorrectly classified instances, Kappa statistics, and mean absolute error, wherein the values obtained are better for neural network than any other classifier. This neural network model would help in accurately predicting theangiographic disease status which is the class attribute indicating percentage of diameter narrowing in diseased patients. This would in turn help in early medical diagnostics.

References

-

How to cite this article

Shabia Shabir Khan and S.M.K. Quadri (2016). Prediction of Angiographic Disease Status using Rule Based Data Mining Techniques , Biological Forum – An International Journal 8(2): 103-107.