geniusnhu

auto_cor = sales.groupby("Date")["Weekly_Sales"].sum()
auto_cor = pd.DataFrame(auto_cor)
auto_cor.columns = ["y"]

# Adding the lag of the target variable from 1 steps back up to 52 (due to a seasonality at the end of the year)
for i in range(1, 53):
    auto_cor["lag_{}".format(i)] = auto_cor.y.shift(i)

# Compute autocorrelation of the series and its lags
lag_corr = auto_cor.corr()

geniusnhu

stepwise_model = pm.auto_arima(Wal_sales.iloc[:,1].values, start_p=1, start_q=1,
                               max_p=20, max_q=20, m=52,
                               start_P=0, seasonal=True,
                               d=1, D=1, trace=True,
                               error_action='ignore',  
                               suppress_warnings=True, 
                               stepwise=True)
print(stepwise_model.aic())

Result

geniusnhu

#Split train and test
train = Wal_sales.iloc[:106,1].values
test = Wal_sales.iloc[106:,1].values
# Train the model
stepwise_model.fit(train)

# Predict test set
pred = stepwise_model.predict(n_periods=37)

# Reframe the data

geniusnhu

# Compute Residual
train_pred = stepwise_model.predict(n_periods=106)
r_train = train - train_pred
r_test = test - pred
residual = pd.DataFrame(np.concatenate((r_train,r_test)), columns={"y"})


# Generate lag of Residuals from 1 step to 52 steps
# Adding the lag of the target variable from 1 steps back up to 52 
for i in range(1, 53):

geniusnhu

# Split train and test sets in correspondence with Time series data
def ts_train_test_split(X, y, test_size):
    test_index = int(len(X)*(1-test_size))
    
    X_train = X.iloc[:test_index]
    y_train = y.iloc[:test_index]
    X_test = X.iloc[test_index:]
    y_test = y.iloc[test_index:]
    
    return X_train, X_test, y_train, y_test

geniusnhu

from tqdm.notebook import tqdm
import plotly.graph_objects as go
import plotly.express as px
import gc
import pandas as pd

from tslearn.clustering import TimeSeriesKMeans
from tslearn.preprocessing import TimeSeriesScalerMeanVariance, TimeSeriesScalerMinMax
from kneed import KneeLocator

geniusnhu

## Example of the main code pipeline in .py format

import sys, os
import pandas as pd
import numpy as np

from your_classes.ClassOne import load_funtion, preprocess_function, training_function, save_result_function
import your_classes.PATH

def main():

geniusnhu

## Example of code annotation for a function

def your_function(X):
    """ Explanation of what the function does 
    
    Parameters
    ----------
    X: dtype
        explanation of X
    y: dtype

geniusnhu

def data_optimize(df, object_option=False):
    """Reduce the size of the input dataframe

    Parameters
    ----------
    df: pd.DataFrame
        input DataFrame
    object_option : bool, default=False
        if true, try to convert object to category

geniusnhu

>>> import sys
>>> my_generator_list = (i for i in range(100000))
>>> print(f"My generator is {sys.getsizeof(my_generator_list)} bytes")
My generator is 128 bytes

>>> timeit(my_generator_list)
10000000 loops, best of 5: 32 ns per loop
  
>>> my_list = [i for i in range(100000)]
>>> print(f"My list is {sys.getsizeof(my_list)} bytes")