Commit Graph

313 Commits (ml_pipeline)

Author SHA1 Message Date
Primoz 26804cf8ea Repair preprocessing one hot encoding of test set. 2023-04-21 13:24:31 +02:00
Primoz 865225994b Added testing section after feature selection. 2023-04-20 13:29:14 +02:00
Primoz 259be708aa Improve the feature selection method with validations etc. 2023-04-20 13:26:20 +02:00
Primoz 0594993133 Add GroupKFold to feature selection CV. Start with generic metric calculation procedure. 2023-04-20 11:20:26 +02:00
Primoz 1cbc743cf7 Add kBest method to initially filter out the worst performing features. Update comments. 2023-04-20 10:12:16 +02:00
Primoz 2a8f1ee613 Merge branch 'ml_pipeline' of https://repo.ijs.si/junoslukan/straw2analysis into ml_pipeline 2023-04-19 15:56:52 +02:00
Primoz ce13a9e13b Implement feature selection method which is used in ML pipeline. 2023-04-19 15:56:34 +02:00
junos 0b16aa6fe4 Clean up categories. 2023-04-18 15:49:33 +02:00
junos d092e17e33 Explore saved categories. 2023-04-18 15:34:06 +02:00
junos d000551884 Ignore only some files in presentation. 2023-04-18 14:57:59 +02:00
junos 5b9a1dba1a Revert "Ignore presentation."
This reverts commit ec7cd09a09.
2023-04-18 14:54:35 +02:00
Primoz 10ca47583c Implement feature selection methods (WIP). 2023-04-14 17:20:22 +02:00
junos f672709ea6 JupyText config. 2023-04-05 20:14:47 +02:00
junos 633d029dc2 Merge remote-tracking branch 'origin/ml_pipeline' into ml_pipeline 2023-03-01 13:37:18 +01:00
junos ec7cd09a09 Ignore presentation. 2023-03-01 13:37:10 +01:00
Primoz 8a532fa95a Add a ML pipeline script to develop a whole pipeline. 2023-02-23 10:41:36 +01:00
Primoz bccc1cd1de Clean and fix Preprocessing module. 2023-02-23 10:40:58 +01:00
Primoz 9ed863b7a1 Add a CrossValidation module with all the required methods. 2023-02-23 10:40:17 +01:00
Primoz f69cb25266 Add planning comments. 2023-02-22 18:12:52 +01:00
Primoz 7f6ae9b323 Add imputation and One-Hot Encoding Methods. 2023-02-22 18:05:01 +01:00
Primoz 8f6cb3f444 Add preprocessing class. 2023-02-22 13:44:03 +01:00
Primoz ef12f64fe5 Add feature selection Class skeleton. 2023-02-20 11:51:34 +01:00
Primoz 63741c3627 Add feature selection pipeline script with initail plan. 2023-02-15 12:27:39 +01:00
Primoz 806ca1b37d Added excel intermediate scores file 2023-02-06 14:02:24 +01:00
Primoz d06da18c26 Fix error with method name. 2023-02-06 11:54:54 +01:00
Primoz 93a34986d9 Comment code sections and change to pd.concatinate method. 2023-02-06 11:31:21 +01:00
Primoz 08e81610a9 Fix method params 2023-02-06 11:21:17 +01:00
Primoz ef78f179cd Add neccessary parameters to plot method. 2023-02-06 11:19:04 +01:00
Primoz afeb7b4872 Remove data_yield from features. 2023-02-06 11:16:53 +01:00
Primoz ea3f805ba7 Change ML model and ddd CV to sequential feat_select. Add std lines to plots. 2023-02-06 11:09:15 +01:00
Primoz e3aef2dae7 Add save to file code, and todo comment 2023-02-01 15:13:57 +01:00
Primoz b286753696 Add vizualization sections for sequential addition of sensors' features. 2023-02-01 13:51:56 +01:00
Primoz 07ef72dec5 Implement algorithm for sequential adding of the most important features. 2023-01-25 14:19:29 +01:00
Primoz 85e572fca0 Expand analysis of the features (individualy and by sensor groups). 2023-01-23 16:32:07 +01:00
Primoz 6a98c8cdcf Prepare scripts for feature importance analysis. 2023-01-19 16:20:43 +01:00
Primoz d263b32564 Temp: remove stratified logo from ml pipeline. 2023-01-19 09:26:55 +01:00
Primoz ad2fab133f Explore features with Entropy and IG. 2023-01-13 17:08:56 +01:00
junos 72fdd9c5ec Use stratified downsampling.
And run all models with a method from machine_learning.helper.
2023-01-04 21:48:37 +01:00
junos b0b9edccc4 Unhide jupyter code cells and outputs. 2023-01-04 21:25:12 +01:00
junos 61d786b2ca Describe LOSO results. 2023-01-04 20:38:05 +01:00
junos 37eada4a2e Present results for stressful events. 2023-01-04 20:00:08 +01:00
junos af6843634c Update RAPIDS submodule. 2023-01-04 18:22:53 +01:00
junos 8bbe0b2ba8 Merge branch 'ml_pipeline' of repo.ijs.si:junoslukan/straw2analysis into ml_pipeline
# Conflicts:
#	.gitignore
#	exploration/ml_pipeline_classification_with_clustering.py
2023-01-04 18:19:43 +01:00
junos 1516d1c000 Cluster by demand_control_ratio_quartile. 2023-01-04 18:16:14 +01:00
Primoz 339142ff31 Add expl stress event script and other changes. 2022-12-21 15:02:25 +01:00
Primoz adcb823d3f Add stress event duration exploration script. 2022-12-15 16:43:40 +01:00
Primoz a61ab9ee51 Add feature importance check. 2022-12-15 16:43:13 +01:00
Primoz 164d12ed2f Add undersampling method (with on/off parameter). 2022-12-13 17:01:46 +01:00
Primoz 0a45e35164 Remove unused imports prt. 2. 2022-12-09 13:56:42 +01:00
Primoz 78b6e7fa07 Remove unused imports from ML pipeline scripts. 2022-12-09 13:53:16 +01:00