Ozbay metric

This is a string metric that measures deviation (or proximity) between 2 words. The metric calculation is a formula that utilizes 3 existing String metric algorithms: Jaccard Distance, Edit Distance and Longest Common Substring Distance.

The formula produces an output value is which is >= 0.0 where 0.0 is a perfect match, and increasing values demonstrate a decreasing likelihood of a match.

Examples

Base word	Term word	Ozbay Metric
ban	ban	0.0
ban	bane	0.3333333
ban	band	0.3333333
ban	bat	0.75
ban	bands	1.3333333
ban	banana	2.0
ban	bandana	2.3333333
ban	bandit	3.0
ban	bandage	4.6666666

This metric accommodates for words that are incorrectly spelt:

Base word	Term word	Ozbay Metric
piccadilly	piccadilly	0.0
piccadilly	piccadilyl	0.25
piccadilly	piccadlily	0.3333333
piccadilly	picacdilly	0.4
piccadilly	picadily	0.5
piccadilly	picacdlily	1.3333333
piccadilly	ipcacdily	1.48148
piccadilly	pcicadlyil	2.0

Python implementation

Executing the implemetation and tests

There is a ozbayMetricRunner.py in the "python/src" folder of the repository. You can supply 2 words as 2 arguments to the file when executing, for it to then calcualte a result and print it out. Here is an example to illustrate:

$ python3 ozbayMetricRunner.py ban bat
The result of comparing ban with bat is: 0.75

To execute the tests in the Python implementation you need to run the "allTests.py" file in the 'python' folder. Before you execute this, you must set the PYTHONPATH to inclue the src folder, like so:

export PYTHONPATH=src

To then execute the file:

$ python3 allTests.py

Using the library in your code

If you want to use the calculator in your source code as a library, You must have a copy of this source code (src folder) in your project. Then in your source code file you need to import the calculator and utilize it declaring it in 2 possible ways:

import ozbayMetricCalculator

You can then utilize it by calling

result = ozbayMetricCalculator.calculate(word1,word2)

Or alternatively:

from ozbayMetricCalculator import calculate 

You can then utilize it by calling

result = calculate(word1,word2)

Java implementation

You must have JDK 1.8 and Maven 3+ installed on your machine in order to run, compile and package the project.

There is a OzbayMetricRunner.java in the 'java/src/main/java/com/hakanozbay/app' folder of the repository.

Executing the implemetation and tests

If you are running the implementation on the command line, without an IDE like Eclipse, you will need to package and compile the project using Maven first, like so:

$ cd java
$ mvn clean package

Afterwards navigate to the built jar pacakage, and supply 2 words as 2 arguments to the file when executing, for it to then calcualte a result and print it out. Here is an example to illustrate:

$ cd target
$ java -cp ozbay-metric-1.0.0.jar com.hakanozbay.app.OzbayMetricRunner ban bat
The result of comparing ban with bat is: 0.75

To execute the tests in the you can provide run the maven test goal, running the "UnitTestSuite" like so:

$ mvn test -Dtest=UnitTestSuite

Using the library in your code

If you want to use the calculator in your source code as a library, assuming you use Maven for you project, you will first need to install the library to your local Maven repository by running the maven install command. You can do this by nagivating to the top level 'java' folder and executing the command:

$ mvn clean install

Afterwards, in your pom.xml file, declare the library as a dependency:

<dependency>
	<groupId>com.hakanozbay</groupId>
	<artifactId>ozbay-metric</artifactId>
	<version>1.0.0</version>
</dependency>

In your source code, you must import the OzbayMetricCalculator declaring it as:

import com.hakanozbay.metric.OzbayMetricCalculator 

You can then utilize it by calling

double result = new OzbayMetricCalculator().calculate(word1,word2)