-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathReadMe.txt
106 lines (58 loc) · 4.2 KB
/
ReadMe.txt
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
# ReadMe file
# Air Liquide : Future Ready Data Challenge.
# Predict My Customer Needs.
# Kasun Bandara, June 2018
======================================================= START ReadMe =====================================================================================================
Folder Structure
================
(*) Scripts folder: Contains all the source files, including the intermediate files generated through scripts().
This also includes the train-dataset.csv, train dataset IM challenge.csv, and test.scv files that are neeed by the initial scripts.
Execusion order:
1) feature_augmentation.R: Creating new feature (combined) using existing timestamp feature.
Input file: train-dataset.csv
Output file:train-features.csv
2) feature_instace_complete.R: generate the incomplete records of each customer.
Input file: train-features.csv
Output file: forecast-complete.csv
3) feature_normalize.R: feature normalization and factorization.
Input file: forecast-complete.csv
Output file: forecast-impute.csv
4) lasso_impute.R: imputing incomplete (N/A) values of Energy consumption using lasso regression.
Input file: forecast-impute.csv
Output file: imputed_timeseries.csv
5) test-dataset.R: Prepares the raw test.csv for the forecasting (Facilitate to match with the appropriate forecast in the forecasting scripts)
Input file: test.csv
Output file: test-forecast.csv
6) forecast-energy-consumption-ets.R: Forecasting using ETS() function from forecast() package.
Input file: imputed_timeseries.csv, test-forecast.csv
Output file: Energy-forecast-ets.csv
7) forecast-energy-consumption-arima.R: Forecasting using Auto.ARIMA() function from forecast() package.
Input file: imputed_timeseries.csv, test-forecast.csv
Output file: Energy-forecast-autoarima.csv
8) forecast-energy-consumption-es.R: Forecasting using ES() function from Smooth() package.
Input file: imputed_timeseries.csv, test-forecast.csv
Output file: Energy-forecast-es.csv
9) ensemble.R: Combining the forecasts using weighting (ensembling)
Input file: Energy-forecast-ets.csv, Energy-forecast-autoarima.csv, Energy-forecast-es.csv
Output file: Kasun-Submission.csv
(*) Files folder: Contains all the source files, including the intermediate files generated through scripts.
1) train dataset IM challenge.csv : Original training datafile given in the competetion.
2) train-dataset.csv : The ordered dataset based on customer (Done using simple Excel filtering)
Note: This file replaces the missing values values of Zip codes by "0" (so we can consider as another category as there are customer IDs with 0 zip codes)
Also removes the record of 1143, as the energy consumption is zero (doesn't add anything for learning), also this customer isn't in the test set.
3) train-features.csv: Contains the new feature (combined) using existing timestamp feature.
4) forecast-complete.csv: Contains complete records of each customer (2015-November to 2018-January)
5) forecast-impute.csv: Normalised records for imputation.
6) imputed_timeseries.csv: Imputed energy consumption values using lasso regression.
7) test.csv: Original test datafile given in the competetion.
8) test-forecast.csv: Prepared test file for forecasting.
9) Energy-forecast-ets.csv: Forecasts of ETS.
10) Energy-forecast-autoarima.csv: Forecasts of ARIMA.
11) Energy-forecast-es.csv: Forecasts of ES.
12) Kasun-Submission.csv : Final submitted file.
(*) Plots folder: Contains all the forecasting plots generated by forecasts.
1) ETS_Forecasts.pdf : Forecasting plot for each customer of test set using ETS()
2) ARIMA_Forecasts.pdf : Forecasting plot for each customer of test set using ARIMA()
3) ES_Forecasts.pdf : Forecasting plot for each customer of test set using ES()
(*) All: Contains all the files (Files in Scripts folder + Files folder+ Plots folder)
======================================================= END ReadMe ============================================================================================================