Download to read offline
Uploaded byAxel de Romblay
Automate Machine Learning Pipeline Using MLBox
The document outlines a hack session led by Axel de Romblay focusing on Automated Machine Learning (Auto-ML) using the MLBox package. It discusses the steps involved in the Auto-ML process including data pre-processing, model tuning, and the automation of a machine learning pipeline aimed at minimizing human intervention while maintaining high performance. It also highlights MLBox's compatibility, ease of installation, and availability of tutorials for users.
More Related Content
Similar to Automate Machine Learning Pipeline Using MLBox
Automate Machine Learning Pipeline Using MLBox
- 1. Hack Session By Axel deRomblay AUTOMATED MACHINE LEARNING
- 2. • Introduction onAuto-ML • MLBox : a powerful Auto-ML python package • Hack session on a dataset AUTOMATED MACHINE LEARNING
- 3. Data ScientistData Computationmeans Data pre-processing Model tuning Machine Learning Almost an automated process…
- 4. Auto Machine Learning Afully automated process Data Computation meansRobot • Supervised tasks - classification - regression • Structured data - csv files - json files - … • Unsupervised tasks - outlier detection - clustering - … • Unstructured data - images - texts - …
- 5. What is auto-ML? We want to automate… …the maximum number of steps in a ML pipeline… …with minimum human intervention… …while conserving a high performance !
- 6. Data cleaning (duplicates, ids, correlations, leaks, …) Data encoding (NA, dates, text, categorical features, … ) STEP 2 : Preprocessing STEP 1 : Reading / merging STEP 3 : Optimisation Feature selection Feature engineering Model selection Prediction Model interpretation STEP 4 : Application Focus on the automation process Diagram of a standard ML pipeline
- 8. Quality: functionalcode : tested on Kaggle Performance: fully distributed and optimised AI: dumping and automatic reading of computations Updates: latest algorithms MLBox: a fully automated python package Compatibility: Python 2.7-3.6, Linux OS Quick setup: $ pip install mlbox User friendly: tutorials, docs, examples…
- 9. Hack Session https://www.kaggle.com/c/two-sigma-connect-rental-listing-inquiries Manual kernel: https://www.kaggle.com/sudalairajkumar/xgb-starter-in-python/ Auto kernel : https://www.kaggle.com/axelderomblay/mlbox-a-fully-automated-package/
- 10.
Editor's Notes
- #2 1min
- #3 1min
- #4 2min Data preprocessing and model tuning are both repetitive tasks that take a lot of time… A Data Scientist is expensive !
- #5 2min So why don’t we replace the DS by a robot ??? We would save time and money ! Let’s see what can be automated !
- #6 1min Performance = computation time + accuracy
- #7 2min 90% of machine learning tasks follow this pipeline
- #8 1min Available on PyPI Github with tutos, examples Docs with articles, kaggle kernels, … Performance : tested on Kaggle ! Features : drifts, embeddings, stacking, leak, feature importances,…
- #9 2min
- #10 40min
- #11 10min THANKS ! Q&A ?