output_of_CNN

This dataset comprises the outcomes obtained by training our convolutional neural network with simulated experimental data sourced from various sensor arrays. The dataset consists of 101 TXT files, where the numbers following "a" in the file names represent the base length of the sensors in the sensor array, the numbers following "h" represent the height of the sensors in the sensor array, and the numbers following "d" represent the distance between the sensor array and the main dipole plane. These numerical values are given in millimeters. A more detailed explanation of these parameters can be found in the maintext. These data include two different formats, which exhibit significant differences in terms of data size. One format has an approximate size of 3KB, where every four rows represent a cycle. In each cycle, the first row describes the model's accuracy in classifying the validation set. The second row corresponds to the first row of the confusion matrix, the third row corresponds to the second row of the confusion matrix, and the fourth row corresponds to the third row of the confusion matrix. The final row after 30 cycles indicates the end of training. The other format has an approximate size of 4000KB, where every six rows represent a cycle. In addition to the previous format, two additional rows are inserted before each cycle. The first row records the training loss for every three training set samples, while the second row records the validation loss for every three validation set samples.