min_value = df[feature_name].min()
result[feature_name] = (df[feature_name] -
min_value) / (max_value - min_value)
return result
for file in data_files:
file_name = "data/" + file + training_suffix
print("Working on: " + file_name + " - started at " + get_date_time())
data = pd.read_csv(file_name)
data[date_column] = data[date_column].apply(transform_date)
data = data[(data[date_column] > 0)]
target = data[target_column]
data.drop(target_column, inplace=True, axis=1)
data.drop(id_column, inplace=True, axis=1)
data.drop(name_column, inplace=True, axis=1)
normalized_data = normalize(data)
ada = md.fit_ada_boost(normalized_data, target, True)
print(ada)
test_data = pd.read_csv("data/" + file + test_suffix)
test_data.drop(name_column, inplace=True, axis=1)
results = pd.DataFrame()
results[id_column] = test_data[id_column]
test_data.drop(id_column, inplace=True, axis=1)
test_data[date_column] = test_data[date_column].apply(transform_date)
normalized_test_data = normalize(test_data)
results[target_column] = ada.predict(normalized_test_data)
results.to_csv(file + "results")
print("end time: " + get_date_time())
import modelling as md
import pandas as pd
import feature_engineering as fe
import os
data = pd.read_csv("data/DJI_1d_10y_signal.csv")
data = data.drop(["Date"], axis=1)
data = data.ffill(axis=0)
print(data)
xTrain, xTest, yTrain, yTest = fe.ordered_train_test_split(data, "Signal")
results = pd.DataFrame()
results["true_y"] = yTest
ada = md.fit_ada_boost(xTrain, yTrain, True)
results["prediction"] = ada.predict(xTest)
print("Adaboost Classifier")
print(results["prediction"])
print(md.get_results(results["true_y"], fe.generate_y(results, "prediction")))
svm = md.fit_SVM(xTrain, yTrain)
results["prediction"] = svm.predict(xTest)
print("SVM Classifier")
print(md.get_results(results["true_y"], results["prediction"]))
print('Accuracy of the SVM on test set: {:.3f}'.format(svm.score(xTest,
yTest)))
knn = md.fit_KNN(xTrain, yTrain, True)
results["prediction"] = knn.predict(xTest)
print("KNN")