2b:null 2c:["$","title",null,{"children":"Fitting Guassians to a Histogram to Identify Peaks | Tanner Hobson | Observable"}] 31:T5ba,# Fitting Guassians to a Histogram to Identify Peaks Suppose you have this histogram plot: ${await PLOT(_0_DATA_ALL)} If you want to identify every peak from this histogram plot, then there's a few methods you could follow. The main idea is that at the end of this process, every peak should be identified and parameterized with a Gaussian function. In this notebook, I identify peaks by simply finding the global maximum: ```javascript ${IDENTIFY_PEAK.toString()} ``` Then, I extract some a small set of x values on either side of that peak: ```javascript ${SUBSET.toString()} ``` Then I can fit a Gaussian function. I am using: ```javascript ${gauss.toString()} ``` After fitting, the fitted Gaussian function can be subtracted from the whole dataset and we can discard bins that are now `<0`. On this dataset, this process works up to 3 iterations, and it does seem to accurately identify peaks for those 3 iterations. The 4th iteration has a bad fit and it throws off the remainder of the results. Overall, these are the peaks identified: ${htl.html`

Solution ${i+1}/${arr.length} (cost = ${cost.toFixed(3)})

Amplitude: ${amplitude.toFixed(3)}
Mean: ${mean.toFixed(3)}
Variance: ${variance.toFixed(3)}