Apache
/
lucene
mirror of https://github.com/apache/lucene.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

SOLR-12701: Improve Monte Carlo example

This commit is contained in:
Joel Bernstein 2018-08-27 20:31:13 -04:00
parent d833b4c9d3
commit 04a50b6a2e
1 changed files with 9 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

15
solr/solr-ref-guide/src/simulations.adoc
View File

@ -46,11 +46,11 @@ The Monte Carlo simulation below performs the following steps:
* A normal distribution with a mean of 2.2 and a standard deviation of .0195 is created to model the length of componentA.
* A normal distribution with a mean of 2.71 and a standard deviation of .0198 is created to model the length of componentB.
* The `monteCarlo` function is used to simulate component pairs. The `monteCarlo` function
calls the *add(sample(componentA), sample(componentB))* function 100000 times and collects the results in an array. Each
time the function is called a random sample is drawn from the componentA
and componentB length distributions. The `add` function adds the two samples to calculate the combined length.
The result of each function run is collected in an array and assigned to the *simresults* variable.
* The `monteCarlo` function samples from the componentA and componentB distributions and sets the values to variables sampleA and sampleB. It then
calls the *add(sampleA, sampleB)* function to find the combined lengths of the samples. The `monteCarlo` function runs a set number of times, 100000 in the example below, and collects the results in an array. Each
time the function is called new samples are drawn from the componentA
and componentB distributions. On each run, the `add` function adds the two samples to calculate the combined length.
The result of each run is collected in an array and assigned to the *simresults* variable.
* An `empiricalDistribution` function is then created from the *simresults* array to model the distribution of the
simulation results.
* Finally, the `cumulativeProbability` function is called on the *simmodel* to determine the cumulative probability
@ -62,7 +62,10 @@ be 5 or less.
----
let(componentA=normalDistribution(2.2, .0195),
componentB=normalDistribution(2.71, .0198),
simresults=monteCarlo(add(sample(componentA), sample(componentB)), 100000),
simresults=monteCarlo(sampleA=sample(componentA),
sampleB=sample(componentB),
add(sampleA, sampleB),
100000),
simmodel=empiricalDistribution(simresults),
prob=cumulativeProbability(simmodel, 5))
----