weight = np.array([car[4] for car in data], dtype=np.float64) acceleration = np.array([car[5] for car in data], dtype=np.float64) model_year = np.array([car[6] for car in data], dtype=np.float64) origin = np.array([car[7] for car in data], dtype=np.float64) corr_vector_to_mpg = [ correlation(mpg, cylinders), correlation(mpg, displacement), correlation(mpg, horsepower), correlation(mpg, weight), correlation(mpg, acceleration), correlation(mpg, model_year) ] abs_corr_vector = [abs(x) for x in corr_vector_to_mpg] weight_to_mpg_m, weight_to_mpg_b = best_fit_line(weight, mpg) cylinder_to_mpg_m, cylinder_to_mpg_b = best_fit_line(cylinders, mpg) displacement_to_mpg_m, displacement_to_mpg_b = best_fit_line(displacement, mpg) horsepower_to_mpg_m, horsepower_to_mpg_b = best_fit_line(horsepower, mpg) acceleration_to_mpg_m, acceleration_to_mpg_b = best_fit_line(acceleration, mpg) model_year_to_mpg_m, model_year_to_mpg_b = best_fit_line(model_year, mpg) mpg_equations = [(cylinder_to_mpg_m, cylinder_to_mpg_b), (displacement_to_mpg_m, displacement_to_mpg_b), (horsepower_to_mpg_m, horsepower_to_mpg_b), (weight_to_mpg_m, weight_to_mpg_b), (acceleration_to_mpg_m, acceleration_to_mpg_b), (model_year_to_mpg_m, model_year_to_mpg_b)] possibilities = find_range_adjustment_results(data, mpg_equations, abs_corr_vector)
mpgData = 'data/auto-mpg.data.txt' dataKey = ['MPG', 'Cylinders', 'Displacement', 'Horsepower', 'Weight', 'Acceleration', 'Model Year'] with open(mpgData) as f: content = f.readlines() data = [[float(x) for x in line.split()[0:7]] for line in content] mpg = np.array([car[0] for car in data], dtype=np.float64) cylinders = np.array([car[1] for car in data], dtype=np.float64) displacement = np.array([car[2] for car in data], dtype=np.float64) horsepower = np.array([car[3] for car in data], dtype=np.float64) weight = np.array([car[4] for car in data], dtype=np.float64) acceleration = np.array([car[5] for car in data], dtype=np.float64) model_year = np.array([car[6] for car in data], dtype=np.float64) weight_to_mpg_m, weight_to_mpg_b = best_fit_line(weight, mpg) cylinder_to_mpg_m, cylinder_to_mpg_b = best_fit_line(cylinders, mpg) displacement_to_mpg_m, displacement_to_mpg_b = best_fit_line(displacement, mpg) horsepower_to_mpg_m, horsepower_to_mpg_b = best_fit_line(horsepower, mpg) acceleration_to_mpg_m, acceleration_to_mpg_b = best_fit_line(acceleration, mpg) model_year_to_mpg_m, model_year_to_mpg_b = best_fit_line(model_year, mpg) mpg_equations = [(cylinder_to_mpg_m, cylinder_to_mpg_b), (displacement_to_mpg_m, displacement_to_mpg_b), (horsepower_to_mpg_m, horsepower_to_mpg_b), (weight_to_mpg_m, weight_to_mpg_b), (acceleration_to_mpg_m, acceleration_to_mpg_b), (model_year_to_mpg_m, model_year_to_mpg_b)] correlation_to_mpg_vector = [correlation(mpg, cylinders), correlation(mpg, displacement), correlation(mpg, horsepower), correlation(mpg, weight), correlation(mpg, acceleration), correlation(mpg, model_year)] abs_correlation_vector = [abs(x) for x in correlation_to_mpg_vector] test_value_vector = [num*(softmax_weight) for num in abs_correlation_vector] weighting_values = softmax(test_value_vector) predicted_mpgs_list = []