Common plotting recipes for statistics and machine learning.
This package uses RecipesBase to provide mappings from types defined in statistics and machine learning packages to generic descriptions of visualization attributes and data. These recipes can then be used in conjunction with Plots.jl to provide flexible statistical and machine learning visualizations which are independent of both the platform and graphical library.
Many recipes are conditionally included and are loaded on the relevant
using call of the library.
using ROCAnalysis, MLPlots will load plotting recipes for roc curves defined in ROCAnalysis. Recipes include:
- OnlineAI: Neural nets and spike trains
- ROCAnalysis: ROC/AUC curves
Status: This package is currently on hiatus until further notice. Collaboration is welcomed and encouraged!
Neural nets with OnlineAI.jl. Show the current state of a neural net:
using OnlineAI, MLPlots net = buildClassificationNet(3, 1, [15,10,5]) plot(net)
For spiking neuron models, a spike (or raster) plot is useful to see firing times among neurons:
n = 20 spikes = SpikeTrains(n, title = "Spike Trains", color = :darkblue) for t=1:100, i=1:n if rand() < 0.1 push!(spikes, i, t) end end spikes.plt
using ROCAnalysis, MLPlots curve = ROCAnalysis.roc(2+2randn(1000), -2+2randn(100000)) plot(curve)