In countries like India, many mortality occurs every year because of improper pronouncement of disease on time. Many people remain deprived of medication as the people per doctor ratio are nearly 1:1700. Every human body and its physiological processes show some symptoms of a diseased condition. The proposed model in this paper would analyze those symptoms for identification of the disease and its type. In this proposed model, few selected attributes would be considered which are shown as symptoms by a person suspected with a particular disease. Those attributes can be taken as input for the proposed symptom analysis and classification model, which is a soft computing model for classifying a sample first to be diseased or disease free and then, if diseased, predicting its type (if any). Number of diseased and disease free samples are to be collected. Each of these samples is a collection of attributes shown / expressed by a human body. With respect to a specific disease, those collected samples form two primary clusters, one is diseased and the other one is disease free. The disease free cluster may be discarded for further analysis. Depending on the symptoms shown by the diseased samples, every disease has some types based on the symptoms it shows. The diseased cluster of samples can reform clusters among themselves depending on the types of the disease. Those clusters then become the classes of the multiclass classifier for analysis of a new incoming sample.
Mishra, Smita Prava Miss; Mishra, Debahuti; and Patnaik, Srikanta Prof.
"A Novel Soft Computing Based Model For Symptom Analysis & Disease Classification,"
International Journal of Pharmacology and Pharmaceutical Technology: Vol. 1
, Article 3.
Available at: https://www.interscience.in/ijppt/vol1/iss1/3