exercise.py

import statsmodels.api as stapi
import numpy

def main():
    (N, X, Y) = read_data()

    results = do_multivariate_regression(N, X, Y)
    print(results.summary())

    effective_variables = get_effective_variables(results)
    print(effective_variables)

def do_multivariate_regression(N, X, Y):
    # 2
    X = numpy.array(X)
    results = stapi.OLS(Y,X).fit()
    print(results)

    return results

def get_effective_variables(results):
    eff_vars = []
	# 3
    n = 1
    for pvalue in results:
        if pvalue < 0.05:
            eff_vars.append(n)
        n+=1

    return eff_vars

def read_data():
    # 1
    N = 0
    X = []
    Y = []
    with open("students.txt") as f:
        next(f)
        for line in f:
            splits = line.strip().split(" ")
            numeric_data = [float(x) for x in splits]
            x = numeric_data[0:-1]
            y = numeric_data[-1]
            X.append(x)
            Y.append(y)
            N += 1

    # X must be numpy.array in (30 * 5) shape.
    # Y must be 1-dimensional numpy.array.
    X = numpy.array(X)
    Y = numpy.array(Y)
    return (N, X, Y)

if __name__ == "__main__":
    main()

Traceback (most recent call last):
File "exercise.py", line 55, in
main()
File "exercise.py", line 10, in main
effective_variables = get_effective_variables(results)
File "exercise.py", line 25, in get_effective_variables
for pvalue in results:
TypeError: 'RegressionResultsWrapper' object is not iterable

def get_effective_variables(results):
eff_vars = []
# 3
n = 0
df_result = pd.DataFrame(results.summary().tables[1].data[1:])

for pvalue in df_result[4]:
    if float(pvalue) < 0.05:
        eff_vars.append(df_result.loc[0][n])
    n+=1
return eff_vars