Fuzzy rule based and anfis classification for rainfall distribution
Throughout history, weather has always been a form of frightful uncertainty to mankind. Other than that, it also has some negative impact on certain industries and jobs including the clothing industry and food production, and it affect peoples’ lives in certain situations especially in terms of maki...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
2012
|
| Subjects: | |
| Online Access: | http://eprints.utm.my/id/eprint/31315/5/NoorHayatiMohdZainMFSKSM2012.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Throughout history, weather has always been a form of frightful uncertainty to mankind. Other than that, it also has some negative impact on certain industries and jobs including the clothing industry and food production, and it affect peoples’ lives in certain situations especially in terms of making travel plans. Hence, it can be said that weather is always related to the daily human activities. Therefore, weather prediction should be classified as an important factor in achieving better living. The data regarding the distribution of weather is usually in a dynamic pattern and hard to predict, which is why the weather classification be done to improve the prediction. The rainfall classification model based on soft computing is feasible to implement and could produce desirable result through the training and testing of the available dataset. Soft computing techniques such as Fuzzy Rule Based (FRB) and Adaptive- Neural Fuzzy Inference System (ANFIS) are investigated in this study to determine which technique is most effective and can achieve higher percentage of accuracy for the purpose of rainfall classification. A number of 720 Senai weather hourly datasets are used in order to test the result of weather classification. The results of the experiments done using both methods show that ANFIS is capable of producing better result for classification with the accuracy percentage of 97.22% in first experiment with FRB which only produce an accuracy of 2.08%. In second experiment, FRB gave 53.47% classification rate, lower than ANFIS which produced higher classifier rate 97.22% thus proving that is it better than FRB. |
|---|