https://github.com/kaltsoon/association-analysis-client

Awesome association analysis client based on FP-Growth algorithm
https://github.com/kaltsoon/association-analysis-client

Last synced: about 2 months ago
JSON representation

Awesome association analysis client based on FP-Growth algorithm

• Host: GitHub
• URL: https://github.com/kaltsoon/association-analysis-client
• Owner: Kaltsoon
• Created: 2017-02-22T15:16:59.000Z (almost 8 years ago)
• Default Branch: master
• Last Pushed: 2017-02-23T09:09:09.000Z (almost 8 years ago)
• Last Synced: 2023-04-06T05:12:19.422Z (almost 2 years ago)
• Language: JavaScript
• Size: 39.1 KB
• Stars: 1
• Watchers: 2
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# Awesome association analysis client based on FP-Growth algorithm

## What?

The algorithm behind the client generates frequent itemsets and generates association rules based on a set of transactions.

Transactions are a single text file containing lines of comma separated items (like shopping basket items). Here's an example

of transaction file:

```

Beer,Diapers

Bread,Diapers

Beer,Butter,Milk

Bread,Milk

Beer,Milk

...

```

The generated rules such as `{ Beer } -> { Milk }`, meaning "in transactions which contain beer also commonly contain milk".

You can read more about assosiation rule learning [here](https://en.wikipedia.org/wiki/Association_rule_learning).

## How?

1. Install [Node.js](https://nodejs.org/en) and [Yarn](https://yarnpkg.com/en/) 

1. Download this repository

2. Run command `yarn` (this command installs the dependencies)

3. Run command `yarn start -- --transactionsFile ./transactions.txt --itemNamesFile ./itemnames.txt --minSupport 0.2 --minConfidence 0.5`

In the 3. step replace `./transactions.txt` with path to transactions file and `./itemnames.txt` with path to item names file. The paths are relative to the folder download in the 2. step.

Transactions file is any file containing list of lines containing comma separated items, like in example above. 

Item names file on the other hand is a file containing all possible items separated by comma. In the given example, item names file would contain:

```

Beer,Diapers,Bread,Milk,Butter

```

The order of the names doesn't matter.

`minSupport` is the minimum support value used to generate frequent itemsets (default is 0.2). `minConfidence` is the minimum confidence value used to generate rules (default is 0.5). If you want to write the output to file set `--outFile` flag value to the output file path. If `outFile` is not provided, output will be written to stdout. If algorithm is executed successfully the output format in `outFile` is the following JSON object:

```

{

  frequentItemsets: FrequentItemset[],

  rules: RuleWithConfidence[]

}

```

where types are following:

```

Itemset = string[];

FrequentItemset = {

  itemset: Itemset,

  support: number,

};

Rule = [Itemset, Itemset];

RuleWithConfidence = {

  rule: Rule,

  confidence: number,

};

```