I'm working with Linked Data (DBpedia in my case) and want to know whether this is the best one among others such as edge counting, information content or maybe hybrid measures.
Maybe these links could be helpful to answer to your question.
http://arxiv.org/ftp/arxiv/papers/1310/1310.8059.pdf
http://www.sciencedirect.com/science/article/pii/S0306457315000023
Hello
There are a lot of approaches for concepts matching (logic, non logic and hybrid)
Logic approaches : based on measures like (Exact, Subsume, Fail)
Non Logic approaches : based on IR techniques for similarity measure (Jiang & Conrath, Lin , Resnik, etc.)
and Hybrid approaches that combines the both approaches for developing more efficient concepts mechanisms.
I hope that help.
If you are working with DBpedia or Wikipedia, you can use word2vec algorithm (https://code.google.com/p/word2vec/). It gives good results for finding semantic similarity between words (concepts).
I recommend you to look random walk/graph- based approachs ( even though most of the time used for word sense disambiguation ). Please get more detail : "From senses to texts: An all-in-one graph-based approach for Measuring Semantic Similarity , MT Pilehvar (2015): its URL of researchgate is "www.researchgate.net/publication/280490105_From_Senses_to_Texts_An_All-in-one_Graph-based_Approach_for_Measuring_Semantic_Similarity"
Article From Senses to Texts: An All-in-one Graph-based Approach for...
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Graphs