# if getting accentedness predictions
# summary_y_values = modeler.get_ys_dict("../../SJP_JC_Audio/perception_results/accented_avgs.csv",
# speaker_list)
# if getting fluency predictions
# summary_y_values = modeler.get_ys_dict("../../SJP_JC_Audio/perception_results/fluency_avgs.csv",
# speaker_list)
# if getting comprehensibility predictions
# summary_y_values = modeler.get_ys_dict("../../SJP_JC_Audio/perception_results/comp_avgs.csv",
# speaker_list)
summary_y_values = modeler.get_ys_dict(fpath, speaker_list)
# zip x and y data
summary_zipped = modeler.zip_feats_and_ys(phono_feats, summary_y_values)
# test to make sure this worked
unzipped_summary_feats, unzipped_summary_ys = zip(*summary_zipped)
# set variables for input into model
unzipped_summary_ys = list(unzipped_summary_ys)
unzipped_summary_feats = np.array(list(unzipped_summary_feats))
summary_shape = unzipped_summary_feats.shape
# create instance of class AdaptiveModel
sum_adapt = modeler.AdaptiveModel(unzipped_summary_feats,
unzipped_summary_ys, summary_shape,
"../../SJP_JC_Audio/phonological_test")
# split data into datasets
"S02", "S03", "S04", "S05", "S07", "S08", "S09", "S19", "S21", "S22",
"S23", "S24", "S25", "S26", "S28"
]
types = [
"../../SJP_JC_Audio/perception_results/accented_avgs.csv",
"../../SJP_JC_Audio/perception_results/fluency_avgs.csv",
"../../SJP_JC_Audio/perception_results/comp_avgs.csv"
]
for fpath in types:
# read in y data
summary_y_values = modeler.get_ys_dict(fpath, speaker_list)
# zip x and y data
summary_zipped = modeler.zip_feats_and_ys(summary_features,
summary_y_values)
# test to make sure this worked
unzipped_summary_feats, unzipped_summary_ys = zip(*summary_zipped)
# set variables for input into model
unzipped_summary_ys = list(unzipped_summary_ys)
unzipped_summary_feats = np.array(list(unzipped_summary_feats))
summary_shape = unzipped_summary_feats.shape
# create instance of class AdaptiveModel
sum_adapt = modeler.AdaptiveModel(unzipped_summary_feats,
unzipped_summary_ys, summary_shape,
dpath)
# split data into datasets
"../data/perception_results/comp_avgs.csv"]
for fpath in y_paths:
# read in y data
# accentedness
y_values = modeler.get_ys_dict(fpath, speaker_list)
# # fluency
# y_values = modeler.get_ys_dict("../../SJP_JC_Audio/perception_results/fluency_avgs.csv",
# speaker_list)
#
# # comprehensibility
# y_values = modeler.get_ys_dict("../../SJP_JC_Audio/perception_results/comp_avgs.csv",
# speaker_list)
# zip x and y data
zipped = modeler.zip_feats_and_ys(acoustic_features, y_values)
# test to make sure this worked
unzipped_feats, unzipped_ys = zip(*zipped)
# set variables for input into model
unzipped_ys = list(unzipped_ys)
unzipped_feats = np.array(list(unzipped_feats))
shape = unzipped_feats.shape
# create instance of class AdaptiveModel
adapt = modeler.AdaptiveModel(unzipped_feats, unzipped_ys, shape, "../audio/IS09_summary")
# split data into datasets
cv_data, cv_ys = adapt.split_data_for_cv(k=10)