Loading...
| Friend's email | |
| Your name | |
| Your email | |
| enter code | |
This page was sent successfuly
- Type of Document: M.Sc. Thesis
- Language: Farsi
- Document No: 40631 (01)
- University: Sharif University of Technology
- Department: Industrial Engineering
- Advisor(s): Akhavan Niaki, Taghi
- Abstract:
- Water demand forecasting and modeling is very important and needful in water resource planning and management as well as water consumption forecasting. The forecasting helps the managers to design and operate various infrastructures of water supply such as tanks and other distribution equipments. Nowadays, intelligent systems are very efficient and practical tools because of their high ability in forecasting and independency from limitative assumptions in classic methods. In this thesis, one of the newest methods, called support vector regression method, is used to forecast monthly demands of water consumption in Tehran, Iran. To develop the method, data is first preprocessed through scaling, smoothing, dimensionality reduction, and clustering methods. The domain of the data is limited by scaling. Then, by generating monthly indices for all variables used in this study, the data is smoothed regarding their monthly fluctuations. Furthermore, by means of stepwise regression, the most effective variables in water consumption prediction are selected. At the end, using the K-means clustering method, the data are grouped and modeled in different clusters. To evaluate the performance of the proposed support vector regression, a feed-forward Perceptron neural networks is employed to forecast the monthly water consumption as well. The results of the comparison study show that the proposed support vector regression method possesses acceptable performance in forecasting the monthly water consumption of Tehran and provides better estimates
- Keywords:
- Pre-Processors ; Smoothing ; Clustering ; Support Vector Regression ; Water Consumption Forecasting
Digital Object List
-
محتواي پايان نامه
- view
Bookmark