def predict_with_v2(forecast, day):
X, y, days = reader.rain_temp_snow(), reader.read_cyclist_data().ravel(), reader.day_of_week()
clf = np.array([linear_model.LinearRegression() for i in range(7)])
for i in range(7):
clf[i].fit(X[days == i], y[days == i])
prediction = clf[day].predict([forecast])
prediction[prediction < 0] = 0
return prediction
def predict_with_v2(forecast, day):
X, y, days = reader.rain_temp_snow(), reader.read_cyclist_data().ravel(
), reader.day_of_week()
clf = np.array([linear_model.LinearRegression() for i in range(7)])
for i in range(7):
clf[i].fit(X[days == i], y[days == i])
prediction = clf[day].predict([forecast])
prediction[prediction < 0] = 0
return prediction
def predict_with_v2b(forecast, day):
X, y, days = reader.rain_temp_snow(), reader.read_cyclist_data().ravel(), reader.day_of_week()
clf = np.array([linear_model.LinearRegression() for i in range(2)])
clf[0].fit(X[days < 5], y[days < 5])
clf[1].fit(X[days > 4], y[days > 4])
if day < 5:
prediction = clf[0].predict([forecast])
else:
prediction = clf[1].predict([forecast])
prediction[prediction < 0] = 0
return prediction
def version1():
X, y = shuffle(reader.rain_temp_snow()[7:], reader.read_cyclist_data().ravel()[7:], random_state=0)
n = int((len(X)/10)*9)
trainingX= X[:n]
trainingY = y[:n]
testX = X[n:]
testY = y[n:]
clf = linear_model.LinearRegression()
clf.fit(trainingX, trainingY)
return clf, testX, testY
def predict_with_v2b(forecast, day):
X, y, days = reader.rain_temp_snow(), reader.read_cyclist_data().ravel(
), reader.day_of_week()
clf = np.array([linear_model.LinearRegression() for i in range(2)])
clf[0].fit(X[days < 5], y[days < 5])
clf[1].fit(X[days > 4], y[days > 4])
if day < 5:
prediction = clf[0].predict([forecast])
else:
prediction = clf[1].predict([forecast])
prediction[prediction < 0] = 0
return prediction
def version1():
X, y = shuffle(reader.rain_temp_snow()[7:],
reader.read_cyclist_data().ravel()[7:],
random_state=0)
n = int((len(X) / 10) * 9)
trainingX = X[:n]
trainingY = y[:n]
testX = X[n:]
testY = y[n:]
clf = linear_model.LinearRegression()
clf.fit(trainingX, trainingY)
return clf, testX, testY
def version2b():
X, y, days = shuffle(reader.rain_temp_snow()[7:],
reader.read_cyclist_data().ravel()[7:],
reader.day_of_week()[7:],
random_state=0)
n = int((len(X) / 10) * 9)
trainingX = X[:n]
trainingY = y[:n]
trainingD = days[:n]
testX = X[n:]
testY = y[n:]
testD = days[n:]
clf = np.array([linear_model.LinearRegression() for i in range(2)])
clf[0].fit(trainingX[trainingD < 5], trainingY[trainingD < 5])
clf[1].fit(trainingX[trainingD > 4], trainingY[trainingD > 4])
return clf, testX, testY, testD
def version2b():
X, y, days = shuffle(
reader.rain_temp_snow()[7:],
reader.read_cyclist_data().ravel()[7:],
reader.day_of_week()[7:],
random_state=0)
n = int((len(X)/10)*9)
trainingX= X[:n]
trainingY = y[:n]
trainingD = days[:n]
testX = X[n:]
testY = y[n:]
testD = days[n:]
clf = np.array([linear_model.LinearRegression() for i in range(2)])
clf[0].fit(trainingX[trainingD < 5], trainingY[trainingD < 5])
clf[1].fit(trainingX[trainingD > 4], trainingY[trainingD > 4])
return clf, testX, testY, testD
