Loading WordVectors as word level features
This notebook will show how to import word vectors into the WordLevelFeatures class.
[1]:
import os
from pyeeg.io import WordLevelFeatures
[2]:
story_id = 0
# Path to data ocntaining onsets of words and duration of speech segments
env_path = '/media/hw2512/SeagateExpansionDrive/EEG_data/Katerina_experiment/story_parts/alignement_data/'
wordfreq_path = '/media/hw2512/SeagateExpansionDrive/EEG_data/Katerina_experiment/story_parts/word_frequencies/'
list_wordfreq_files = [item for item in os.listdir(wordfreq_path) if item.endswith('timed.csv')]
list_stories = [item.strip('_word_freq_timed.csv') for item in list_wordfreq_files]
list_env_files = [os.path.join(env_path, s, s + '_125Hz.Env') for s in list_stories]
# Path to WordVectors
wv_path = '/home/hw2512/MachineLearning/Word2Vec/GloVe-1.2/vectors_correct_header.txt'
# Loading word onset and duration for one story:
wo_path = os.path.join(wordfreq_path, list_wordfreq_files[story_id])
duration_path = os.path.join(env_path, list_env_files[story_id])
[3]:
wf = WordLevelFeatures(path_praat_env=duration_path, path_wordonsets=wo_path, path_wordvectors=wv_path)
INFO:pyeeg.io:Word flushington not in word embedding model; will use rdm instead
INFO:pyeeg.io:Word bashfulest not in word embedding model; will use rdm instead
INFO:pyeeg.io:Word adulation not in word embedding model; will use rdm instead
INFO:pyeeg.io:Word blush not in word embedding model; will use rdm instead
INFO:pyeeg.io:Word blush not in word embedding model; will use rdm instead
INFO:pyeeg.io:Word paucity not in word embedding model; will use rdm instead
INFO:pyeeg.io:Word fellow-men not in word embedding model; will use rdm instead
INFO:pyeeg.io:Word fastened not in word embedding model; will use rdm instead
INFO:pyeeg.io:Word flushington not in word embedding model; will use rdm instead
INFO:pyeeg.io:Word scraggy not in word embedding model; will use rdm instead
INFO:pyeeg.io:Word timid not in word embedding model; will use rdm instead
INFO:pyeeg.io:Word deprecating not in word embedding model; will use rdm instead
INFO:pyeeg.io:Word uninteresting not in word embedding model; will use rdm instead
INFO:pyeeg.io:Word impervious not in word embedding model; will use rdm instead
INFO:pyeeg.io:Word aimless not in word embedding model; will use rdm instead
INFO:pyeeg.io:Word superstitiously not in word embedding model; will use rdm instead
INFO:pyeeg.io:Word monosyllabic not in word embedding model; will use rdm instead
INFO:pyeeg.io:Word laboriously not in word embedding model; will use rdm instead
INFO:pyeeg.io:Word flushington not in word embedding model; will use rdm instead
INFO:pyeeg.io:Word gyp not in word embedding model; will use rdm instead
INFO:pyeeg.io:Word bedmaker not in word embedding model; will use rdm instead
INFO:pyeeg.io:Word unaccountably not in word embedding model; will use rdm instead
INFO:pyeeg.io:Word sluices not in word embedding model; will use rdm instead
INFO:pyeeg.io:Word flushington not in word embedding model; will use rdm instead
[19]:
feat_matrix = wf.align_word_features(features=['wordvectors'], srate=100, wordonset_feature=True)
print("First few samples of first 4 dimensions:\n")
print(feat_matrix[250:350, :4])
First few samples of first 4 dimensions:
[[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 1. 0.07879453 0.09271022 -0.1146835 ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 1. -0.74636799 0.196638 0.490275 ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]
[ 0. 0. 0. 0. ]]