선택한 단어 수는 입니다.
최소 단어 이상 선택하여야 합니다.
최소 단어 이상 선택하여야 합니다.
선택한 단어 수는 30입니다.
최대 10 단어까지만 선택 가능합니다.
최대 10 단어까지만 선택 가능합니다.
제안 방법
- 0, for the source power, bias power, and gas ratio, respectively. Additional six experiments were conducted to provide the test data for model evaluation. Consequently, a total of 15 experiments were conducted.
- As an illustration, the Al etch rate was modeled. First, the effects of three training factors (TT, NHN, IWD) were optimized, followed by the optimization of two gradients. The experimental ranges for TT, NHN, IWD employed in the first stage are 0.
- In this study, a BPNN model was constructed by controlling the training and test errors during a genetic learning. This technique was evaluated with the experimental data collected from the etching of silica thin films [7].
- In this study, the effect of training factors on the BPNN prediction performance was optimized by using GA. GA played a role of finding optimized training factor sets as a function of trading factor between the training and test errors.
- To characterize the etch process, a 23 full factorial experiment [8] was used along with one center point. The experimental parameter ranges employed in the design are 100-800 W, 100-400 W and 0.2-5.0, for the source power, bias power, and gas ratio, respectively. Additional six experiments were conducted to provide the test data for model evaluation.
대상 데이터
- 2. At each generation, the constructed 200 models were evaluated and only one model with the smallest P-RMSE was selected. This was repeated for 100 generations and one finally optimized model was then determined among 100 selected models.
데이터처리
- 6, respectively. The prediction error was quantified with the root mean squared error (RMSE). For convenience, this error is called "P-RMSE”.
이론/모형
- A buffer-clad layer of about 25 um was deposited by the flame hydrolysis deposition method. The core layer of 8 um thickness was subsequently formed, which was formed as SiO2- P2O5-B2O3-GeO2.
※ AI-Helper는 부적절한 답변을 할 수 있습니다.