Journal of Applied Optics, Volume. 43, Issue 6, 1037(2022)

Prediction model of K2CsSb photocathode reflectivity based on LSTM

Jingwen WEI... Yunsheng QIAN* and Yang CAO |Show fewer author(s)
Author Affiliations
  • School of Electronic and Optical Engineering, Nanjing University of Science and Technology, Nanjing 210094, China
  • show less
    Figures & Tables(13)
    Structure model of K2CsSb photocathode
    Structure diagram of reflectivity prediction model
    Thermodynamic diagram of dataset feature
    Diagram of dataset serialization reconstruction
    LSTM calculation principle and unit structure diagram
    Structure diagram of reflectivity monitoring system
    Trend chart of sequence length, accuracy and training time
    Comparison of loss values under different layers of LSTM network
    Comparison of model prediction effects under different datasets
    • Table 1. Composition of network structure and the number of parameters

      View table
      View in Article

      Table 1. Composition of network structure and the number of parameters

      网络层输出维度参数量
      输入层(None, 5, 5)0
      LSTM(None, 5, 10)640
      随机失活层(None, 5, 10)0
      LSTM(None, 10)840
      随机失活层(None, 10)0
      输出层(None, 1)11
    • Table 2. Parameters of experimental model

      View table
      View in Article

      Table 2. Parameters of experimental model

      主要参数参数值
      损失函数MSE
      学习率0.0003
      优化器Adam
      序列长度9
      网络层数2
      迭代次数100
      批处理大小/条32
    • Table 3. Effect of sequence length on accuracy and training time

      View table
      View in Article

      Table 3. Effect of sequence length on accuracy and training time

      序列 长度每轮训练 时间/s准确率/%序列 长度每轮训练 时间/s准确率/%
      12.1398.421115.2198.45
      34.3698.631315.4698.26
      55.1299.211515.7998.91
      75.6398.821715.8698.99
      99.6299.011915.9898.43
    • Table 4. Effect of network layers on accuracy and training time

      View table
      View in Article

      Table 4. Effect of network layers on accuracy and training time

      BP神经网络LSTM网络
      网络 层数每轮训练 时间/s准确率/%网络 层数每轮训练 时间/s准确率/%
      12.1290.8913.2698.79
      24.5294.4525.1299.21
      35.8995.6936.6598.93
    Tools

    Get Citation

    Copy Citation Text

    Jingwen WEI, Yunsheng QIAN, Yang CAO. Prediction model of K2CsSb photocathode reflectivity based on LSTM[J]. Journal of Applied Optics, 2022, 43(6): 1037

    Download Citation

    EndNote(RIS)BibTexPlain Text
    Save article for my favorites
    Paper Information

    Category: Research Articles

    Received: Aug. 8, 2022

    Accepted: --

    Published Online: Nov. 18, 2022

    The Author Email:

    DOI:10.5768/JAO202243.0604001

    Topics