DISEASE PREDICTION SYSTEM USING DATA MINING
- 1. DISEASE PREDICTION SYSTEM USING DATA MINING Under Guidance of asst. prof. Ashutosh Pandey Presented By:- Anand kumar mishra (1616210020) Siddhesh shukla (1616210112) Shivani yadav (1616210103) B.Tech. (CS) – 4th year.
- 2. OUTLINE • Introduction ▪ Objective ▪ Motivation • Literature Survey • Some Key Features of Disease • Plan of Action • Methodology Adopted ▪ Data Collection ▪ Steps to be Performed ▪ Functional Architecture • Expected Result • Conclusion • References
- 3. OBJECTIVE Identifying hidden patterns and relationships among various attributes that can lead to: ▪ better diagnosis, ▪ better medicines, ▪ better treatment ▪ Early diagnosis may predict the chances of Disease and lead to take preventive measures before the situation becomes critical.
- 4. MOTIVATION • The prevalence of Diabetes is increasing in all countries and its prevention has become a public health priority. • The predictors of Diabetes risk are insufficiently understood.
- 5. WORLDWIDE STATISTICS ❏ The number of people with diabetes has risen from 108 million in 1980 to 422 million in 2014. ❏ The global prevalence of diabetes among adults over 18 years of age has risen from 4.7% in 1980 to 8.5% in 2014 ❏ Diabetes is a major cause of blindness, kidney failure, heart attacks, stroke and lower limb amputation. ❏ In 2016, an estimated 1.6 million deaths were directly caused by diabetes. ❏ Almost half of all deaths attributable to high blood glucose occur before the age of 70 years.
- 7. INDIAN STATISTICS � Diabetes currently affects more than 62 million Indians, which is more than 7.1% of the adult population. � Nearly 1 million Indians die due to diabetes every year. TOTAL DEATHS (in million) SEX % DEATH DUE TO DIABETES 40.2 MALE 8 23.8 FEMALE 12
- 8. MOTIVATION ❖ Recent research has shown that the onset of can be postponed or prevented with lifestyle intervention or by medication. ❖ Identifying individuals at high risk of cancer has therefore become a priority for targeting preventive measures effectively. ❖ Symptoms are often less marked, thus the disease may be diagnosed several years after onset, once complications have already arisen
- 9. SYMPTOMS OF DIABETES � increased urine output, � excessive thirst, � weight loss, � hunger, � fatigue, � skin problems � slow healing wounds, � yeast infections, and � tingling or numbness in the feet or toes.
- 11. PLAN OF WORK Data Collection / Data Analysis & Literature Review (3-4 months) Data Warehouse Construction (1-2 months) Building Classifier (3-3.5 months) Experimental Result Analysis (1-2 months) Report Preparation (1-1.5 months)
- 12. METHODOLOGY ADOPTED Step 1- Preprocessing of data Step 2-Dividing the patients into different group Step 3- Apply the Fuzzy Inference Step 4-Using Apriori Algorithms to find the relative pattern Step 5- Build Classifier
- 13. FUNCTIONAL ARCHITECTURE DATA SET PREPROCESSING ASSOCIATION RULE MINNING FUZZIFICATION OF ATTRIBUTES RESULT BUILD CLASSIFIER DIVIDE THE DATA SET INTO 3 CLUSTERS
- 14. PREPROCESSING -- MERGING DATA FROM MULTIPLE SOURCES INTO UNIQUE FORMAT -- MISSING VALUE HANDLING Use the attribute mean for all samples belonging to the same as the given tuple.
- 15. DIVIDE THE DATA SET INTO 3 CLUSTER Dividing the patients into 3 three different group according to different condition of patients � Very serious � Serious � Normal
- 16. FUZZY LOGIC � Intuitionistic Fuzzy Set : claim that an element x belongs to a given degree μA(x) to a fuzzy set x should not belong to A to the extent 1- μA(x) α - Cut: Let α be a number between 0 and 1. The α-cut of fuzzy set A at level α is the set of those elements of A where membership function is greater than or equal to α. 16
- 17. APRIORI ALGORITHM We use apriori algorithm to find relative pattern Suppose we get association rule A→B confidence 95% Means if patient has attribute A then it will has attribute B also with confidence of 95 %
- 18. BUILD CLASSIFIER � Various techniques may be applied – ✔ Multilayer Back propagation feed-forward ANN -- Train ANN using the weight of attributes calculate from association rules confidence value. ✔ Simple Weighted Sum Method � Expected Result : � -- Class Label prediction of the patient as either: � normal � serious or � very serious ✔
- 19. EXPECTED RESULT � Class Label prediction of the patient as either: � normal � serious or � very serious o Association among various attributes with respective confidence level (A🡪B , CL)
- 20. CONCLUSION We are finding the relative pattern of patients in a hospital we uses IFS, α-cuts, and Apriori algorithm for discovering the knowledge of patients. Our approach will successfully protect the patients’ personal data privacy and will achieve some gratifying results from the experiments. Certainly, the approach is not limited in a disease, it can be used in other fields in the long run.
- 21. REFERENCES • “Mining Cancer data with Discrete Particle Swarm Optimization and Rule Pruning “ Yao Liu and Yuk Ying Chun • “Identifying HotSpots in Lung Cancer Data Using Association Rule Mining “ Ankit Agrawal and Alok Choudhary • “Comparison of feature selection methods for multiclass cancer classification based on microarray data” Xiaobo Li1,2*, Sihua Peng3, Xiaosi Zhan1 • “Lung cancer statistics,” centers for Disease Control and Prevention, URL:http://www.cdc.gov/cancer/lung/statistics • en.wikipedia.org/wiki/World_Health_Organization • www.whoindia.org • A. Jemal, F. Bray, M.M. Center, J. Ferlay, E. Ward, D. Forman(2011). "Global cancer tatistics". CA: a cancer journal forclinicians61