Created
June 4, 2024 16:14
-
-
Save preetesh33/b2447be10dfe4c3e5db53d369961b48f to your computer and use it in GitHub Desktop.
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
import pandas as pd | |
from mlforecast import MLForecast | |
from xgboost import XGBRegressor | |
from window_ops.rolling import rolling_mean, rolling_max, rolling_min, rolling_std | |
from mlforecast.forecast import MLForecast | |
from mlforecast import MLForecast | |
import os | |
from joblib import dump | |
import json | |
# Suppressing the warning | |
import warnings | |
warnings.filterwarnings("ignore", category=FutureWarning) | |
warnings.filterwarnings("ignore", category=DeprecationWarning) | |
organized_file = "/opt/ML/final/data_augmentation/augmentation/merged_main.csv" | |
df_grouped = pd.read_csv(organized_file) | |
# print(df_grouped.head(10)) | |
df_grouped['timestamp'] = pd.to_datetime(df_grouped['timestamp']) | |
# Convert object columns to category type | |
df_grouped['location'] = df_grouped['location'].astype('category') | |
df_grouped['customer'] = df_grouped['customer'].astype('category') | |
df_grouped['cluster'] = df_grouped['cluster'].astype('category') | |
df_grouped['project'] = df_grouped['project'].astype('category') | |
# Create a mapping of codes to labels for each category column dictionary | |
location_mapping_dict = dict(enumerate(df_grouped['location'].cat.categories)) | |
customer_mapping_dict = dict(enumerate(df_grouped['customer'].cat.categories)) | |
cluster_mapping_dict = dict(enumerate(df_grouped['cluster'].cat.categories)) | |
project_mapping_dict = dict(enumerate(df_grouped['project'].cat.categories)) | |
################################# Save the mapping dictionaries as json files ############################## | |
with open('location_mapping.json', 'w') as f: | |
json.dump(location_mapping_dict, f, indent=4) | |
with open('customer_mapping.json', 'w') as f: | |
json.dump(customer_mapping_dict, f, indent=4) | |
with open('cluster_mapping.json', 'w') as f: | |
json.dump(cluster_mapping_dict, f, indent=4) | |
with open('project_mapping.json', 'w') as f: | |
json.dump(project_mapping_dict, f, indent=4) | |
########################################################################################################### | |
# Convert category columns to integer encoding | |
df_grouped['location'] = df_grouped['location'].cat.codes | |
df_grouped['customer'] = df_grouped['customer'].cat.codes | |
df_grouped['cluster'] = df_grouped['cluster'].cat.codes | |
df_grouped['project'] = df_grouped['project'].cat.codes | |
# Split data into train and valid sets | |
train=df_grouped[df_grouped['timestamp']<='2024-02-15'] #y/m/d | |
# Create an ID column for each unique ID in the training data | |
train['id_col'] = train['location'].astype(str) + '_' + train['customer'].astype(str) + '_' + train['cluster'].astype(str) + '_' + train['project'].astype(str) | |
print("Data split in train set") | |
print("train head",train.head(10)) | |
################################################## Training ################################################### | |
model=XGBRegressor(random_state=990, n_estimators=500, learning_rate=0.01, max_depth=10, reg_lambda=0.2) | |
print("part1") | |
fcst=MLForecast(model, freq='min',lags=[60], lag_transforms={1:[(rolling_mean, 60),(rolling_max, 60),(rolling_min, 60),(rolling_std, 60)]}, | |
date_features=['day','hour','minute'],num_threads=48) | |
print("part2") | |
fcst.fit(train, id_col='id_col', time_col='timestamp', target_col='totalcalls', static_features=['location','customer','cluster','project']) | |
print("part3") | |
######################## Save model ################################################## | |
print("Saving model") | |
dump(fcst, 'test_train_1.joblib') | |
#################################################################################### |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment