TODO #1
Description
TODO
- Important
- Decision Trees
- Check Niambh Bayes
- Experiment on different feature sets and compare with recommendations
- Covariance categorical: http://www.suri.cs.okayama-u.ac.jp/~niitsuma/e/covEigGiniRep17.pdf
- categorical intuition: https://www.youtube.com/watch?v=zLHunbpH5Hg
- Extra
- Try Logistic Classifier
- Neural Network
- Report
- Main insights from contingency tables
- If LaTeX copy benjis format: https://www.benjaminrosman.com/papers/comshons07.pdf
- Give reasons for basis functions chosen, decisions, justify methods
- Try different basis functions maybe
- Compare models with scikit Learn implementations - discuss differences
- Go through "Applying Machine Learning" tut for report
- Visualisations and graphs
- contingency tables
- maybe bar graphs : see
- Very useful (for non categorical/one-hot encoded datasets) https://machinelearningmastery.com/quick-and-dirty-data-analysis-with-pandas/
- Scatter plot matrix: https://plot.ly/python/scatterplot-matrix/
- ggparallel thing in R: https://github.com/heike/ggparallel
- MCA python package: http://vxy10.github.io/2016/06/10/intro-MCA/ and https://github.com/esafak/mca/blob/master/docs/usage.rst
- Better MCA package: https://github.com/MaxHalford/Prince
- Parallel coordinate plot https://stackoverflow.com/a/16907551
- Parallel Coordinate plot nice http://benalexkeen.com/parallel-coordinates-in-matplotlib/
- Use http://webdemo.myscript.com/views/math.html# for quick LaTeX formulas
- Mention using decision tree as PCA and for doing dimensionality reduction
- Mention our results vs suggested results/ variables to use , that the data set suggested
- Try doing the tree or other thing without odor
- Write "use your crayons" on plots
- Mushroom diagrams and stuff: https://www.usask.ca/biology/fungi/glossary.html