# mathisonian/kde

Fetching contributors…
Cannot retrieve contributors at this time
124 lines (96 sloc) 5.01 KB
 [meta title:"Kernel Density Estimation" description:"A useful statistical tool that sounds scarier than it is." twitterHandle:"mathisonian" shareImageUrl:"https://mathisonian.github.io/kde/images/share.png" shareImageWidth:"1280" shareImageHeight:"640" /] [var name:"state" value:"title" /] [var name:"bandwidth" value:0.05 /] [var name:"kernel" value:"epanechnikov" /] [var name:"amplitude" value:3.0 /] [Fixed fullWidth:true ] [KDEVisualizer state:state bandwidth:bandwidth kernel:kernel amplitude:amplitude /] [/Fixed] [TextContainer] [Scroller currentState:state] [Step state:"title" className:"header"] [Header title:"Kernel Density Estimation" author:"Matthew Conlen" authorLink:"https://twitter.com/mathisonian" /] [/Step] [Step state:"start-drop"] Kernel density estimation is a really useful statistical tool with an intimidating name. Often shortened to **KDE**, it's a technique that let's you create a smooth curve given a set of data. This can be useful if you want to visualize just the "shape" of some data, as a kind of continuous replacement for the discrete histogram. It can also be used to generate points that *look like they came from a certain dataset* - this behavior can power simple simulations, where simulated objects are modeled off of real data. I hope this article provides some intuition for how KDE works. [/Step] [Step state:"show-generator"] To understand how KDE is used in practice, lets start with some points. The white circles on your screen were sampled from some unknown distribution. As more points build up, their silhouette will roughly correspond to that distribution, however we have no way of knowing its true value. [/Step] [Step state:"show-estimate"] The [Blue]blue line[/Blue] shows an estimate of the underlying distribution, this is what KDE produces. The KDE algorithm takes a parameter, *bandwidth*, that affects how "smooth" the resulting curve is. Use the control below to modify bandwidth, and notice how the estimate changes. Bandwidth: [Dynamic value:bandwidth min:0.001 max:0.1 step:0.001 /] [/Step] [Step state:"build-estimate"] The KDE is calculated by weighting the distances of all the data points we've seen for each location on the [Blue]blue line[/Blue]. If we've seen more points nearby, the estimate is higher, indicating that probability of seeing a point at that location. Move your mouse over the graphic to see how the data points contribute to the estimation — the "brighter" a selection is, the more likely that location is. The [Red]red curve[/Red] indicates how the point distances are weighted, and is called the *kernel function*. The points are colored according to this function. *Click to lock the kernel function to a particular location.* [/Step] [Step] Changing the bandwidth changes the shape of the kernel: a lower bandwidth means only points very close to the current position are given any weight, which leads to the [Blue]estimate[/Blue] looking squiggly; a higher bandwidth means a shallow kernel where distant points can contribute. Bandwidth: [Dynamic value:bandwidth min:0.001 max:0.1 step:0.001 /] Next we'll see how different kernel functions affect the estimate. [/Step] [Step] The concept of weighting the distances of our observations from a particular point, [equation latex:"x" /], can be expressed mathematically as follows: [equation display:true] \hat{f}(x) = \sum_{observations}^{}{K(\frac{x - observation}{bandwidth})} [/equation] The variable [equation latex:"K" /] represents the kernel function. Using different kernel functions will produce different estimates. Use the dropdown to see how changing the kernel affects the estimate. Kernel: [Select value:kernel options:[ {label:"Epanechnikov", value:"epanechnikov"}, {label: "Normal", value: "normal"}, {label: "Uniform", value: "uniform"}, {label: "Triangular", value: "triangular"} ] /] [br/] Bandwidth: [Dynamic value:bandwidth min:0.001 max:10 step:0.001 /] [br/] Amplitude: [Dynamic value:amplitude min:0.1 max:10.0 step:0.1 /] [/Step] [Step] That's all for now, thanks for reading! I'll be making more of these quick explainer posts, so if you have an idea for a concept you'd like to see, [link url:"https://twitter.com/mathisonian" text:"reach out on twitter"/]. Here are a few useful links: [ul] [li][link text:"Idyll: the software used to write this post" href:"https://idyll-lang.org/" /].[/li] [li][link text:"The source code for this post" href:"https://github.com/mathisonian/kde/" /].[/li] [li][link text:"Learn more about kernel density estimation" href:"https://en.wikipedia.org/wiki/Kernel_density_estimation" /].[/li] [li][link text:"A D3 + JavaScript KDE reference" href:"https://bl.ocks.org/mbostock/4341954" /].[/li] [/ul] [/Step] [/Scroller] [/TextContainer] [analytics google:"UA-108267630-1" /]
You can’t perform that action at this time.