def train_shuffle():
df_train, labels = data_preprocess.process(38, shuffle_index=True, index_file=0)
model_gene_based.run_model_kfold(df_train, labels, epochs, index=0)
df_train, labels = data_preprocess.process(38, shuffle_index=True, index_file=1)
model_gene_based.run_model_kfold(df_train, labels, epochs, index=1)
df_train, labels = data_preprocess.process(38, shuffle_index=True, index_file=2)
model_gene_based.run_model_kfold(df_train, labels, epochs, index=2)
df_train, labels = data_preprocess.process(38, shuffle_index=True, index_file=3)
model_gene_based.run_model_kfold(df_train, labels, epochs, index=3)
def train():
# deepAMR_run.main()
df_train, labels = data_preprocess.process(
38, gene_dataset=True) # SNP or GENE
# gene_dataset_creator.main()
# df_train, labels = data_preprocess.process(38)
# source_analysis.main(df_train)
# df_train, labels = data_preprocess.process(38, shuffle_operon_group=True)
# ML_methods.model_run(df_train, labels) # BO for ML and GBT
# model_gene_based.run_model(df_train, labels, epochs)
# wide_n_deep.run_bayesian(df_train, labels) # BO for wide n deep method
# model_gene_based.run_bayesian(df_train, labels) # main diff in data folding nested CV # init_point 15, n_iter 15 # run_one_fold also run_k_fold and run_single_fold (section.3.5 help accuracy)
# model_gene_based.run_bayesian_single(df_train, labels)
# model_gene_based.run_all(df_train, labels, epochs)
model_gene_based.run_model_kfold(df_train, labels, epochs)
for i in range(len(y) - 1, -1, -1):
if y[i][0] != 0.0 and y[i][0] != 1.0:
del y[i]
del X[i]
y = to_categorical(y)
for i in range(0, len(X)):
if len(X[i]) < ((len(X[i]) // FrameSize + 1) * FrameSize):
for j in range(
0,
(((len(X[i]) // FrameSize + 1) * FrameSize) - len(X[i]))):
X[i].append(0)
X[i] = np.reshape(X[i], (FrameSize, len(X[i]) // FrameSize))
X = np.array(X)
y = np.array(y)
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.1)
model_CNN256_LSTM128_64_2(FrameSize, X, X_train, X_test, y_train, y_test,
epoch)
# model_256_128_64_2BS(FrameSize, X, X_train, X_test, y_train, y_test, epoch)
model_256_128_64_2(FrameSize, X, X_train, X_test, y_train, y_test, epoch)
if __name__ == '__main__':
df_train, labels = data_preprocess.process(6)
run_model(df_train, labels, 20)
def main():
"""Running the LRCN model on Imperial HPC"""
df_train, labels = data_preprocess.process(38, gene_dataset=True)
epochs = 200
model_gene_based.run_model_kfold(df_train, labels, epochs)
import pandas as pd
from loading_data import data_preprocess
def LoadLabel(path):
dt = pd.read_csv(path + 'AllLabels' + '.csv')
dt.set_index(dt.columns[0], inplace=True, drop=True)
return dt
df_train, dt = data_preprocess.process(2, 0)
dfPyrazinamide = dt[0]
arr = dfPyrazinamide.values.tolist()
zero = 0
one = 0
nan = 0
for i in range(0, len(arr)):
if arr[i][0] == 1:
one = one + 1
elif arr[i][0] == 0:
zero = zero + 1
else:
nan = nan + 1
print("streptomycin")
print(one)
print(zero)
print(nan)
def train():
# deepAMR_run.main()
df_train, labels = data_preprocess.process(38, gene_dataset=True)
# gene_dataset_creator.main()
# df_train, labels = data_preprocess.process(38)
# source_analysis.main(df_train)
# df_train, labels = data_preprocess.process(38, shuffle_operon_group=True)
# ML_methods.model_run(df_train, labels)
# model_gene_based.run_model(df_train, labels, epochs)
# wide_n_deep.run_bayesian(df_train, labels)
# model_gene_based.run_bayesian(df_train, labels, comp_iter=1)
# model_gene_based.run_bayesian_kfold(df_train, labels, comp_iter=1)
# model_gene_based.run_bayesian_single(df_train, labels)
# model_gene_based.run_all(df_train, labels, epochs)
# model_gene_based.run_model_kfold(df_train,labels,epochs)
# base_approach.run_feature_importance(df_train, labels)
# base_approach.find_feature_importance(file_name='feature_importance/score_results/lime_all.csv', k=200)
# base_approach.find_feature_importance(file_name='feature_importance/score_results/feature_scores_lime_all_test.csv', k=200)
# base_approach.find_feature_importance(file_name='feature_importance/score_results/feature_scores_lime_test_200.csv', k=200)
# base_approach.find_feature_importance(file_name='feature_importance/score_results/feature_scores_lime_train.csv', k=200)
#
# print("______")
# base_approach.find_feature_importance(file_name='feature_importance/score_results/lime_all.csv', k=100)
# base_approach.find_feature_importance(file_name='feature_importance/score_results/feature_scores_lime_all_test.csv',
# k=100)
# base_approach.find_feature_importance(file_name='feature_importance/score_results/feature_scores_lime_test_200.csv',
# k=100)
# base_approach.find_feature_importance(file_name='feature_importance/score_results/feature_scores_lime_train.csv',
# k=100)
# print("______")
# base_approach.find_feature_importance(file_name='feature_importance/score_results/lime_all.csv', k=50)
# base_approach.find_feature_importance(file_name='feature_importance/score_results/feature_scores_lime_all_test.csv',
# k=50)
# base_approach.find_feature_importance(file_name='feature_importance/score_results/feature_scores_lime_test_200.csv',
# k=50)
# base_approach.find_feature_importance(file_name='feature_importance/score_results/feature_scores_lime_train.csv',
# k=50)
# print("______")
# base_approach.find_feature_importance(file_name='feature_importance/score_results/lime_all.csv', k=20)
# base_approach.find_feature_importance(file_name='feature_importance/score_results/feature_scores_lime_all_test.csv',
# k=20)
# base_approach.find_feature_importance(file_name='feature_importance/score_results/feature_scores_lime_test_200.csv',
# k=20)
# base_approach.find_feature_importance(file_name='feature_importance/score_results/feature_scores_lime_train.csv',
# k=20)
# for i in range(1, 8):
# print(i)
# base_approach.find_feature_importance(file_name='feature_importance/score_results/split/feature_scores_lime_train_150_' + str(i) + '.csv', k=200)
#
# for i in range(1, 8):
# print(i)
# base_approach.find_feature_importance(file_name='feature_importance/score_results/split/feature_scores_lime_train_150_' + str(i) + '.csv', k=100)
#
# for i in range(1, 8):
# print(i)
# base_approach.find_feature_importance(file_name='feature_importance/score_results/split/feature_scores_lime_train_150_' + str(i) + '.csv', k=50)
#
# for i in range(1, 8):
# print(i)
# base_approach.find_feature_importance(file_name='feature_importance/score_results/split/feature_scores_lime_train_150_' + str(i) + '.csv', k=20)
i2 = []
for i11 in range(0, 25):
i2.append(i11)
print(i2)
with multiprocessing.Pool(processes=25) as pool:
pool.map(lime.main_function, i2)
def train():
df_train, labels = data_preprocess.process(2)
print(type(df_train))
models.model_one_vs_all.run_model(df_train, labels, epochs, limited=True)
ML_methods.model_run(df_train, labels)