Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/justinshenk/learnthestructure
Implementation of libpgm for the Wisconsin Breast Cancer Database
https://github.com/justinshenk/learnthestructure
bayesian-network python
Last synced: 27 days ago
JSON representation
Implementation of libpgm for the Wisconsin Breast Cancer Database
- Host: GitHub
- URL: https://github.com/justinshenk/learnthestructure
- Owner: JustinShenk
- Created: 2016-10-06T12:42:09.000Z (over 8 years ago)
- Default Branch: master
- Last Pushed: 2016-10-09T23:41:13.000Z (over 8 years ago)
- Last Synced: 2024-11-07T11:53:50.330Z (3 months ago)
- Topics: bayesian-network, python
- Language: Python
- Size: 202 KB
- Stars: 1
- Watchers: 3
- Forks: 1
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Bayesian Network Structure Learning with Breast Cancer Data
Implementation of [libpgm](https://github.com/CyberPoint/libpgm) to estimate the Bayesian Network structure of the [Wisconsin Breast Cancer Database](https://archive.ics.uci.edu/ml/datasets/Breast+Cancer+Wisconsin+(Original)). This code was written by Marc Vidal and Justin Shenk for Nico Potyka's "Basic Methods in Probabilistic Reasoning" seminar.
# Project structure
Python script is located in `libpgm-1.3/implementation/learnthestructure.py`.
The original data is located in `libpgm-1.3/data/breast-cancer-wisconsin.data`.
# Getting started
Clone the repository
`git clone https://github.com/JustinShenk/learnthestructure.git`
Install libpgm package
`cd learnthestructure`
`cd libpgm-1.3`
`[sudo] python setup.py install`
Run the script
`cd implementation`
`python learnthestructure.py [p-value parameter] [# of bins] [lg]`
The data is discrete rather than continuous. To test the binning of the linear Gaussian function, however, add `lg` as an argument and specify the number of bins. By default, p-value threshold is .05 and linear Gaussian data is discretized into 10 bins.
# Data variables
| *#* | Attribute | Domain |
| --- | --- | --- |
| 1. | Sample code number | id number |
| 2. | Clump Thickness | 1 - 10 |
| 3. | Uniformity of Cell Size | 1 - 10 |
| 4. | Uniformity of Cell Shape | 1 - 10 |
| 5. | Marginal Adhesion | 1 - 10 |
| 6. | Single Epithelial Cell Size | 1 - 10 |
| 7. | Bare Nuclei | 1 - 10 |
| 8. | Bland Chromatin | 1 - 10 |
| 9. | Normal Nucleoli | 1 - 10 |
| 10. | Mitoses | 1 - 10 |
| 11. | Class: | (2 for benign, 4 for malignant) |
# Output
This implementation outputs estimated edges and vertices from the data in files marked with optional arguments:
`/data/breast-data-result-0.1-10.txt`
CPDs are produced:
`/data/breast-data-result-CPDs.txt`
Example output is at the bottom.
# Query
Open the python shell from the implementation folder and instantiate the class:
`from learnthestructure import LearnTheStructure`
`bn = LearnTheStructure()`
What is the probability that a patient's cancer is malignant given that Bare Nuclei has a value of 10?
`evidence = dict(BareNuclei=10)`
`query = dict(Class=[4])`
`bn.query_it(evidence,query)`
`The probability of {'Class': [4]} given {'BareNuclei': 10} is 0.808146976884`
```
Edges:
[
[
"UniformityofCellShape",
"UniformityofCellSize"
],
[
"UniformityofCellSize",
"Class"
],
[
"BareNuclei",
"Class"
]
]
```
```
Vertices data with CPD:
{
"SingleEpithelialCellSize": {
"vals": [
2,
7,
3,
1,
6,
4,
5,
8,
10,
9
],
"numoutcomes": 10,
"cprob": [
0.5522174535050072,
0.017167381974248927,
0.10300429184549356,
0.06723891273247497,
0.058655221745350504,
0.06866952789699571,
0.055793991416309016,
0.030042918454935622,
0.044349070100143065,
0.002861230329041488
],
"parents": [],
"children": []
},
"UniformityofCellSize": {
"vals": [
1,
4,
8,
10,
2,
3,
7,
5,
6,
9
],
"numoutcomes": 10,
"cprob": [
0.5493562231759657,
0.05722460658082976,
0.04148783977110158,
0.09585121602288985,
0.06437768240343347,
0.07439198855507868,
0.027181688125894134,
0.04291845493562232,
0.03862660944206009,
0.008583690987124463
],
"parents": [],
"children": [
"UniformityofCellShape",
"Class"
]
},
"BareNuclei": {
"vals": [
1,
10,
2,
4,
3,
9,
7,
5,
8,
6
],
"numoutcomes": 10,
"cprob": [
0.597997138769671,
0.1888412017167382,
0.04291845493562232,
0.027181688125894134,
0.04005722460658083,
0.012875536480686695,
0.011444921316165951,
0.04291845493562232,
0.030042918454935622,
0.005722460658082976
],
"parents": [],
"children": [
"Class"
]
},
"UniformityofCellShape": {
"vals": [
1,
4,
8,
10,
2,
3,
5,
6,
7,
9
],
"numoutcomes": 10,
"cprob": {
"['1']": [
0.8619791666666666,
0.013020833333333334,
0.0,
0.0,
0.06770833333333333,
0.057291666666666664,
0.0,
0.0,
0.0,
0.0
],
"['8']": [
0.0,
0.06896551724137931,
0.3793103448275862,
0.06896551724137931,
0.0,
0.034482758620689655,
0.0,
0.06896551724137931,
0.27586206896551724,
0.10344827586206896
],
"['3']": [
0.17307692307692307,
0.15384615384615385,
0.019230769230769232,
0.0,
0.21153846153846154,
0.25,
0.1346153846153846,
0.057692307692307696,
0.0,
0.0
],
"['6']": [
0.0,
0.2222222222222222,
0.037037037037037035,
0.037037037037037035,
0.0,
0.07407407407407407,
0.14814814814814814,
0.3333333333333333,
0.1111111111111111,
0.037037037037037035
],
"['10']": [
0.0,
0.029850746268656716,
0.08955223880597014,
0.7164179104477612,
0.014925373134328358,
0.029850746268656716,
0.029850746268656716,
0.029850746268656716,
0.04477611940298507,
0.014925373134328358
],
"['7']": [
0.0,
0.10526315789473684,
0.2631578947368421,
0.15789473684210525,
0.0,
0.0,
0.05263157894736842,
0.05263157894736842,
0.3157894736842105,
0.05263157894736842
], ...
}
```