A class to represent a Bayes Network. And apply all the different sampling strategies to the calculate a given query.
Project description
Bayes Nets and Sampling
About:
Developed code from scratch to compute an input probability query on a given Bayes net on discrete random variables using Prior sampling, Rejection Sampling, Likelihood weighting and Gibbs sampling.
I had also written helper function to generate samples from any input univariate discrete distribution and then use it in your Bayes Net sampling code.
BayesNets Class:
I implemented a BayesNets Class to do all the required task and computations.
Basically I created a single class to do all the calculation, extractions of details from the file as well as implement all the sampling methods and sample generation.
Install the Python Package
pip install bayes-nets-sample
Import the Python Package
from bayes import bayesNets
To see the Sample .txt file
from bayes.sample_input_txt import SampleInputTXT
SampleInputTXT().generate_sample()
To Get more information about the BayesNets Class as well as its methods:
run the following command:
help(BayesNets)
Lets Start By Creating an object of BayesNets and extract the information from the file:
bayesNet = BayesNets(filepath = "example_bayesnet.txt")
Let's Print The Extracted Information:
print(bayesNet)
Let's See how a single sample is generated:
bayesNet.generateSample()
Now Lets Calculate the Query result using Prior Sampling Method:
With verbose = True
bayesNet.doPriorSampling(TOTAL_SAMPLE = 10000, verbose = True)
Now Lets Calculate the Query result using Prior Sampling Method:
With verbose = False
- Just to see the final result.
bayesNet.doPriorSampling(TOTAL_SAMPLE = 10000, verbose = False)
Now Lets Calculate the Query result using Rejection Sampling Method:
With verbose = True
bayesNet.doRejectionSampling(TOTAL_SAMPLE_REQUIRED = 10000, verbose = True)
Now Lets Calculate the Query result using Rejection Sampling Method:
With verbose = False
- Just to see the final result.
bayesNet.doRejectionSampling(TOTAL_SAMPLE_REQUIRED = 10000, verbose = False)
Let's See how a single weighted sample is generated:
bayesNet.generateWeightedSample(verbose = True)
Now Lets Calculate the Query result using Likelihood Weighting Method:
With verbose = True
bayesNet.doLikelihoodWeighting(TOTAL_SAMPLE = 10000, verbose = True)
Now Lets Calculate the Query result using Likelihood Weighting Method:
With verbose = False
- Just to see the final result.
bayesNet.doLikelihoodWeighting(TOTAL_SAMPLE = 10000, verbose = False)
Sample File
5, B, E, A, J, M
B, +b, -b
E, +e, -e
A, +a, -a
J, +j, -j
M, +m, -m
B |
+b, 0.001
-b, 0.999
E |
+e, 0.002
-e, 0.998
A | B,E
+b, +e, +a, 0.95
+b, +e, -a, 0.05
+b, -e, +a, 0.94
+b, -e, -a, 0.06
-b, +e, +a, 0.29
-b, +e, -a, 0.71
-b, -e, +a, 0.001
-b, -e, -a, 0.999
J | A
+a, +j, 0.9
+a, -j, 0.1
-a, +j, 0.05
-a, -j, 0.95
M | A
+a, +m, 0.7
+a, -m, 0.3
-a, +m, 0.01
-a, -m, 0.99
Query: P( B=+b| J=+j)
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Hashes for bayes_nets_sample-1.4-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 13928a074c8d4c062cb8cd3e9b141e1937c29fbc520950a54155ec8e11e13eea |
|
MD5 | fac560b38f947ba50ffc9c4d3c469546 |
|
BLAKE2b-256 | 79903a8a22df4eb95ebd7d6bf0eb37fe56ca078d8750ad0e91fd3be78a6a0d11 |