Dice.com tutorial on using black box optimization algorithms to tune your Solr search engine configuration, by Simon Hughes ( Dice Data Scientist ). See 'Automated Relevancy Tuning using Black Box Optimization Algorithms.ipynb' for Jupyter Notebook tutorial on using black box optimization algorithms from sci-kit optimize to tune your solr config. Example fake Dice jobs data and a solr core are provided for this example. They are not representative of our real data, nor of our actual Solr search implementation.
Solr (Solr core is for 5.0+, for earlier solr versions or later ones, you will need to modify the configs and re-index the data from the notebook).
Python 2.7
Python libraries:
- solrpy
- pandas
- numpy
- scikit-optimize
- https://github.com/mitre/quaerite - uses a GA to optimize Solr parameters