from os import path
import requests
import numpy as np
import pandas as pd
from time_series_model_basics import moving_average, smoothing, metrics, optimizers
from statsmodels.tsa.api import ExponentialSmoothing, SimpleExpSmoothing, Holt

Load Data

own_col = 'home_made_holt'
ts_col = 'sales'
nb_replicate = 10*2+1

datafile = '../data/superstore_sales.csv'

df = pd.read_csv(
    datafile,
    index_col='Order Date',
    dtype={
        'Row ID': str,
        'Order ID': str,
    },
    parse_dates=['Order Date', 'Ship Date'],
)

df.columns = ['_'.join(x.lower().split(' ')) for x in df.columns]

df = pd.concat(nb_replicate * [df.reset_index()[[ts_col]]])
df.index = pd.date_range('1700-01-01', periods=len(df), freq='D')

df1 = df.copy()
df2 = df.copy()

smoothing_level = 0.1
smoothing_trend = 0.1

%%time
stats_holt = Holt(
    df1[[ts_col]],
    initialization_method="estimated",
).fit(
    smoothing_level=smoothing_level,
    smoothing_trend=smoothing_trend,
    optimized=False,
)
sm_holt = 'sm_holt'
df1[sm_holt] = stats_holt.fittedvalues

CPU times: user 1.99 s, sys: 178 ms, total: 2.17 s
Wall time: 1.84 s

%%time 

_ = smoothing.DOUBLE(
    [smoothing_level, smoothing_trend],
    df=df2[[ts_col]],
    ts_col=ts_col,
)

CPU times: user 7.85 s, sys: 30.7 ms, total: 7.88 s
Wall time: 7.88 s

Title

Load Data