Parameter Automatic Calibration Approach of CyclonicPrecipitation Forecast Models Using a Neural Network
| Autor: | Tsai, En-Ping, 蔡恩平 |
|---|---|
| Rok vydání: | 2015 |
| Druh dokumentu: | 學位論文 ; thesis |
| Popis: | 103 This paper presents a parameter automatic calibration (PAC) approach of cyclonic precipitation forecast models using artificial neural network (ANN). A classical ANN-based model, the multilayer perceptron (MLP) neural network, was adopted to demonstrate the feasibility and accuracy of the proposed ANN–PAC approach. In this study, the learning rate, momentum, and number of neurons in the hidden layer were used in the design code as its major parameters. The Dawu gauge station in Taitung, Taiwan, was the study site, and observed typhoon characteristics and ground weather data were the study data. The results obtained of ANN–PAC model is verified with the benchmark solution of the traditional multiple linear regression model. The ANN–PAC model was more reliable than those yielded by the trial-and-error calibration method. In this thesis, the computing efficiency of the ANN–PAC model decreased with an increase in the number of increments within the parameter ranges because of the considerably increased computational time, whereas the prediction errors decreased because of the model’s increased capability of identifying optimal solutions. |
| Databáze: | Networked Digital Library of Theses & Dissertations |
| Externí odkaz: |
|