scores = cross_val_score(clf, X_new, y_2d, cv=5, scoring='accuracy')
print "accuracy scores with 5 fold cross validation with reduced features for ", to_predict, scores
print "mean of accuracy", scores.mean()
if __name__ == '__main__':
# raw_path_string = raw_input("Enter path where data is located (Location of accession number dirs): ")
# csv_path = raw_input("Enter path of directory to store csv files: ")
# train_path = raw_input("Enter path of train csv file (Path upto p1_train.csv): ")
slash = "/"
raw_path_string = '/home/rasika/Documents/Computational Biology/Project/Data'
csv_path = '/home/rasika/Documents/Computational Biology/Project/Result'
train_path = '/home/rasika/Documents/Computational Biology/Project/p1_train_pop_lab.csv'
# make csv files from quant.sf files
make_csv.make_csv_files(raw_path_string + slash, csv_path, slash)
colnames1 = ['TPM']
classifier_input = list()
label_dict = {}
# store the labels from train file in a dictionary
train_data = pd.read_csv(train_path, sep=',', header=0, dtype='unicode')
for i, row in train_data.iterrows():
label_dict[row[0]] = (row[1], row[2])
classifier_input = list()
print "Starting reading csv files"
print datetime.datetime.now()
print "accuracy scores with 5 fold cross validation for Population DT", scores
print "mean of accuracy", scores.mean()
if __name__ == '__main__':
raw_path_string = raw_input("Enter path where data is located (Location of accession number dirs): ")
csv_path = raw_input("Enter path of directory to store csv files: ")
train_path = raw_input("Enter path of train csv file (Path upto p1_train.csv): ")
slash = "\\"
# raw_path_string = '/home/rasika/Documents/Computational Biology/Project/Data'
# csv_path = '/home/rasika/Documents/Computational Biology/Project/Result'
# train_path = '/home/rasika/Documents/Computational Biology/Project/p1_train_pop_lab.csv'
# make csv files from quant.sf files
colnames1 = ['TPM','Length']
make_csv.make_csv_files(raw_path_string + slash, csv_path, slash, ['Name'] + colnames1)
classifier_input = list()
label_dict = {}
# store the labels from train file in a dictionary
train_data = pd.read_csv(train_path, sep=',', header=0, dtype='unicode')
for i, row in train_data.iterrows():
label_dict[row[0]] = (row[1], row[2])
classifier_input = list()
print "Starting reading csv files"
print datetime.datetime.now()
files = listdir(csv_path)
def main():
args = sys.argv
global model_dump_path
global slash
global csv_path
global train_path
model_dump_path = args[1]
raw_path_string = args[2]
csv_path = args[3]
train_path = args[4]
slash = args[5]
eq_class = args[6]
dataframe_csv_path = args[7]
scores = {}
df = None
# w_file = open('/home/rasika/Documents/Computational Biology/Project/output_file.txt', 'a')
# w_file.write('\n\n' + str(datetime.datetime.now()))
col_names = ['TPM', 'Length']
# make csv files from quant.sf files
make_csv.make_csv_files(raw_path_string + slash, csv_path, slash,
['Name'] + col_names)
label_dict = {}
# store the labels from train file in a dictionary
train_data = pd.read_csv(train_path, sep=',', header=0, dtype='unicode')
for i, row in train_data.iterrows():
label_dict[row[0]] = (row[1], row[2])
print "Started reading csv files"
print datetime.datetime.now()
# Reading the data from csv files and creating a data list of acession number, tpm, length and effective length
files = listdir(csv_path)
if eq_class == 'False':
df = create_dataframe(files, ['TPM', 'Length'], col_names, label_dict)
else:
df = parse_eq_classes.create_dataframe_with_eq_class(
raw_path_string, csv_path, train_path, slash, label_dict,
dataframe_csv_path)
print "Read all csv files, created dataframe"
print datetime.datetime.now()
# created dataframe will have following format
# Name TPM_1 TPM_2 TPM_3 TPM_4 .... TPM_199324 label
# ERR188021 value value value value .... value TSI
# ERR188022 . . . . .... . CEU
# . . . . . .... . .
# . . . . . .... . .
# . . . . . .... . .
# w_file.write("\n Predicting for Population, Sequence Center and Both on full data: TPM, Length, Effective Length, NumReads\n")
f1_score_p, accuracy_p = predict_population(df)
scores['Population'] = (f1_score_p, accuracy_p)
f1_score_sc, accuracy_sc = predict_sequence_center(df)
scores['Sequence Center'] = (f1_score_sc, accuracy_sc)
f1_score_p_sc, accuracy_p_sc = predict_population_seq_center(df)
scores['Population and Sequence Center'] = (f1_score_p_sc, accuracy_p_sc)
return scores