https://github.com/kensuke-mitsuzawa/strength-of-association
A simple python library of Strength of Association
https://github.com/kensuke-mitsuzawa/strength-of-association
Last synced: about 2 months ago
JSON representation
A simple python library of Strength of Association
- Host: GitHub
- URL: https://github.com/kensuke-mitsuzawa/strength-of-association
- Owner: Kensuke-Mitsuzawa
- Created: 2015-09-02T12:00:48.000Z (over 9 years ago)
- Default Branch: master
- Last Pushed: 2015-09-02T12:02:31.000Z (over 9 years ago)
- Last Synced: 2025-01-19T11:30:17.665Z (4 months ago)
- Language: Python
- Size: 141 KB
- Stars: 0
- Watchers: 2
- Forks: 0
- Open Issues: 1
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# What's this？
This is simple package of SoA(Strength of Association).
With SoA, you can get (label, word, Association_score).
SoA formula is following Mohammad Saif M, Kiritchenko Svetlana, "Using Hashtags to Capture Fine Emotion Categories from Tweets", 2013
# Setting up
This library is confirmed to work under python 2.7.3
This library has no dependencies.
```
[sudo] python setup.py install
```
# Usage
## input data
input data must be dict structure.
Key is label name. Value is 2-dimension list. like `[ ["word"] ]`
For example,
```
{
"label_a": [
["I", "aa", "aa", "aa", "aa", "aa"],
["bb", "aa", "aa", "aa", "aa", "aa"],
["I", "aa", "hero", "some", "ok", "aa"]
],
"label_b": [
["bb", "bb", "bb"],
["bb", "bb", "bb"],
["hero", "ok", "bb"],
["hero", "cc", "bb"],
],
"label_c": [
["cc", "cc", "cc"],
["cc", "cc", "bb"],
["xx", "xx", "cc"],
["aa", "xx", "cc"],
]
}
```
## how to use
### initialize
`n_thread` is optional argument. If you don't specify, SoA is calculated with single thread.
```
n_thread = 5
soa_obj = SoaClass(labeled_corpus=input_dict, n_thread=n_thread)
```
### calculate
`calc_soa` method returns SoA calculated score.
`is_Zero` is optional. Default is `False`
If you give `is_Zero==True`, you can get all word including score is Zero.
```
soa_score = soa_obj.calc_soa(is_zero=True)
```
### output format
output data is dict type. Key is label name and Value is list.
`{'word': unicode or string, 'score': float}` inside value list.
```
{'label_a': [
{'score': 5.0, 'word': 'aa'}, {'score': 1.4150374992788437, 'word': 'ok'}, {'score': 0.4150374992788437, 'word': 'hero'}, {'score': -1.7548875021634687, 'word': 'bb'}
],
'label_b': [
{'score': 2.807354922057604, 'word': 'bb'}, {'score': 1.8073549220576042, 'word': 'hero'}, {'score': 0.8073549220576041, 'word': 'ok'}, {'score': -2.0, 'word': 'cc'}
],
'label_c': [
{'score': 3.6147098441152083, 'word': 'cc'}, {'score': -2.3625700793847084, 'word': 'bb'}, {'score': -2.777607578663552, 'word': 'aa'}
]}
```