# imports
from EEG_Familiarity import preprocData Filter / Trail Selections
Select specific trails and labels for the problem of binary classifications
To prepare the data into a machine learning model ready format, we need to do the following preprocessing steps. Along the steps, we need to provide some information regarding the positive and negative class.
Specified the file path, and instantiate a preproc object.
file_path = "../data/data_CRMN_vs_MMN_imbalLDA_order_proj_1.mat"
data_preproc = preproc(file_path, experiment_num=1)
data_preproc<EEG_Familiarity.preproc.preproc>Specify the positive class and negative class index via preproc.filter_index. For the numbering system, please refer to Data Format
pos1, neg1 = data_preproc.filter_index(2,5,2,4)
pos2, neg2 = data_preproc.filter_index(4,5,4,4)Based on the filter, we can do an inner merge operation between two class using preproc.merge_two_class. After merging, we can get the data directly using preproc.get_data_by_index.
pos_idx, neg_idx = data_preproc.merge_two_class(pos1, neg1, pos2, neg2)
X, y, subject = data_preproc.get_data_by_index(pos_idx, neg_idx)By doing so, we constructed the \(\mathbb{X}\), \(\mathbb{y}\) from the data of this specific classifier, along with a subject identifiers.
X.shape, y.shape, subject.shape((3813, 72), (3813,), (3813,))From the output, this particular classifier has \(3813\) observation and \(72\) dimensional features associate with them.