High-Definition-Research-Tech-Challenge

#Lower Capacity Strategy

1. Review the fields in a meeting and discuss which are likelier to be easier to clean.

2. Ignore the harder fields and focus only on the easier ones. Build the simplest possible classifier (usually logistic regression or decision tree) on these fields.

3. Focus next on producing summary statistics for this subset of fields, and some basic charts/ graphs.

4. Collate into a notebook.

5. Add commentary and interpretation but acknowledge the limitations of not considering all the fields.

6. Prepare into a simple slide deck and practice timed delivery.

7. If you accomplish the above quicker than expected you can go back over to:

• clean and incorporate some of the omitted fields
• rerun the model with these, or consider alternative models
• expand your presentation – perhaps include more impressive visuals
• do further background reading to improve your understanding of the data

Insert visualisations from data exploration and ML prediction plots

Useful links

Trello board: https://trello.com/invite/b/t28LcStz/ATTIa9253884ed88829cf8c54d13d7bb910c893D98C1/group10-technical-challenge

Machine Learning Algorithms Cheatsheet [Python/R]: https://www.kaggle.com/discussions/getting-started/156497

Extensive overview of ML with Caret: https://www.machinelearningplus.com/machine-learning/caret-package/#:~:text=Caret%20Package%20is%20a%20comprehensive%20framework%20for%20building,the%20optimal%20model%20in%20the%20shortest%20possible%20time.

Feature selection with Caret: https://machinelearningmastery.com/feature-selection-with-the-caret-r-package/

ML Ensembles: https://machinelearningmastery.com/machine-learning-ensembles-with-r/