Dataset for Digital Twin paper

Description for each folder:

1) The "raw_signal_data" folder contains original sensor data. The origigital sensor data is named as "ACF-x-x.xlsx". In each excel document, three directions of vibration, current and force sensor data is included.  Naming rules is as follows: "ACF-1-2.xlsx" represents the orginal sensor data from first layler's second slot in. "ACF-2-3.xlsx" represents the orginal sensor data from second layler's third slot.

2) The original real measured surface roughness is in the "Ra.xlsx". The corresponding cutting parameters for every raw signal data of each slot are shown in "Ra.xlsx" document as well. 
3) The prediction results from different models are save in "prediction_results". The detalies are listed below.

   "BPNN_all_signal.xlsx" is the prediction result from BPNN used the combination of all sensors data as the model input 
    "BPNN_current_signal.xlsx" is the prediction result from BPNN used only the current sensor data as the model input 
    "BPNN_force_signal.xlsx" is the prediction results from BPNN used only the force sensor data as the model input 
    "BPNN_vibration_signal.xlsx" is the prediction result from BPNN used only the vibration sensor data as the model input
    "ELM.xlsx" is the prediction result from ELM used the combination of all sensors data as the model input 
    "GPR.xlsx" is the prediction result from GPR used the combination of all sensors data as the model input
    "LASSO.xlsx" is the prediction result from LASSO used the combination of all sensors data as the model input
    "MLR.xlsx" is the prediction result from MLR used the combination of all sensors data as the model input
    "SVR.xlsx" is the prediction result from SVR used the combination of all sensors data as the model input
    "cnn_all_signal.xlsx" is the prediction result from CNN used the combination of all sensors data as the model input 
    "cnn_current.xlsx" is the prediction result from CNN used only the current sensor data as the model input 
    "cnn_force.xlsx" is the prediction results from CNN used only the force sensor data as the model input 
    "cnn_vibration.xlsx" is the prediction result from CNN used only the vibration sensor data as the model input
    "real_ytest.xlsx" is the real-measured surface roughness data for testing models.

各文件夹说明如下：
1) "raw_signal_data" 文件夹存储原始传感器数据，该类数据文件命名格式为"ACF-x-x.xlsx"。每份Excel文件均包含振动、电流及力三类传感器的实测数据。命名规则如下："ACF-1-2.xlsx"代表第1层第2个槽位的原始传感器数据；"ACF-2-3.xlsx"代表第2层第3个槽位的原始传感器数据。
2) 实测原始表面粗糙度数据存储于"Ra.xlsx"文件中，该文件同时记录了每个槽位对应原始信号数据的切削参数。
3) 不同模型的预测结果存储于"prediction_results"文件夹中，具体说明如下：
"BPNN_all_signal.xlsx"为采用全传感器数据作为模型输入的反向传播神经网络（Back Propagation Neural Network, BPNN）预测结果
"BPNN_current_signal.xlsx"为仅采用电流传感器数据作为模型输入的BPNN预测结果
"BPNN_force_signal.xlsx"为仅采用力传感器数据作为模型输入的BPNN预测结果
"BPNN_vibration_signal.xlsx"为仅采用振动传感器数据作为模型输入的BPNN预测结果
"ELM.xlsx"为采用全传感器数据作为模型输入的极限学习机（Extreme Learning Machine, ELM）预测结果
"GPR.xlsx"为采用全传感器数据作为模型输入的高斯过程回归（Gaussian Process Regression, GPR）预测结果
"LASSO.xlsx"为采用全传感器数据作为模型输入的套索回归（Least Absolute Shrinkage and Selection Operator, LASSO）预测结果
"MLR.xlsx"为采用全传感器数据作为模型输入的多元线性回归（Multiple Linear Regression, MLR）预测结果
"SVR.xlsx"为采用全传感器数据作为模型输入的支持向量回归（Support Vector Regression, SVR）预测结果
"cnn_all_signal.xlsx"为采用全传感器数据作为模型输入的卷积神经网络（Convolutional Neural Network, CNN）预测结果
"cnn_current.xlsx"为仅采用电流传感器数据作为模型输入的CNN预测结果
"cnn_force.xlsx"为仅采用力传感器数据作为模型输入的CNN预测结果
"cnn_vibration.xlsx"为仅采用振动传感器数据作为模型输入的CNN预测结果
"real_ytest.xlsx"为用于模型测试的实测表面粗糙度基准数据。