Replication Data for:"Real-World Considerations for Deep Learning in Wireless Signal Identification Based on Spectral Correlation Function"

The dataset includes spectral correlation function (SCF) estimations by FFT accumulation method (FAM) for totally 4500 signals with 20000 I/Q samples. The signals belong to three different cellular communication standards: GSM, WCDMA, and LTE. The signals have been received from the different channels with multipath, fading, and noise. The dataset can be used to validate the designed classifier model aiming to identify cellular communication signals. For each signal, the dimension of SCF estimate is 8193*16. There are two train sets which must be used together (concatenate train_data_wo_mapping1 and train_data_wo_mapping2 ). Two train sets have 3000 signals totally, and the test set has 1500. The label of the cellular communication standards are given in dataset as follows: WCDMA -> 0 LTE -> 1 GSM -> 2 The dataset includes: 1. SCFDatatrain1.mat 2. SCFDatatrain2.mat 3. SCFDatatest.mat The contents of .mat files: train_class : denotes class labels of the train set, its dimension is 3000*1 double train_data_wo_mapping1 : includes the first half of the training data, its dimension 1500*1 cell train_data_wo_mapping2 : includes the second half of the training data, its dimension 1500*1 cell *Note, concatenate two cells given above (ie [train_data_wo_mapping1; train_data_wo_mapping2]) test_class : denotes class labels of the train set, its dimension is 1500*1 double test_data_without_mapping : includes the test data, its dimension 1500*1 cell Each cell contains 1500 SCF estimates (8193*16) . The dataset has been used for the paper "Real-World Considerations for Deep Learning in Wireless Signal Identification Based on Spectral Correlation Function" submitted for possible publication in IEEE Wireless Communication Letters. Please cite this paper, if you use the dataset.