This project is a collection of various string similiarity algorithms implemented in Java and conveniently packaged into a single jar file with no dependencies.
- Hamming Distance
- Jaro Distance
- Jaro-Winkler Distance
- Dice Coefficient Metric
- Levenshtein Distance
- Damerau-Levenshtein Distance
- Needleman Wunsch Distance
- Smith Waterman Distance
- Optimal String Alignment Distance
- Gotoh Distance
- Ratcliff/Oberhshelp String Matching
- Soundex Coding
The following metrics are vector based metrics. The strings are represented as frequency count vectors (bag of words model).
- L1 Distance
- L2 Distance
- Cosine Distance
- Jaccard Distance
- Dice Coefficient
- Matching Coefficient
- Overlap Coefficient
- QGram Metric
- Tversky Index
The project comes in form of a single jar (plus a jar containing the source code and another jar for the javadoc, if you are interested).
To get a working jar you can:
- Download one of the prepackaged versions
- Clone this repository and use
mvn packageto generate a jar from the latest source - Add the project as a maven dependency
to come
TODO: briefly describe architecture (string and vector metrics, interfaces, default getter, builders) TODO: code example for one metric
The project is released under the Apache 2.0 license. In plain language that means that you can modify and distribute it and that commercial use is allowed. We only ask you to mention us somewhere in your project.