1 of 5
A service to calculate the semantic similarity between a pair of sentence.
Started as a final year group project, after deep research and chunk wise development resulted in production ready application.
LexiContext is my Bachelor of Engineering final year project.
The idea behind the project is that often we as a human being can understand and determine how much similar two sentences are, but for a machine to understand and determine the similarity percentage is not that simple as it sounds.
We started researching for this project from the later phase of 3rd year i.e 5th Semester and we referred dozens of research papers.
Up until recently we came up with an approach with satisfactory result and thought of optimizing the algorithm.
Today we have released our product as "Semantic Text Analytics as a service" which is in beta phase.
Natural Language Processing is one of the evolving fields given the ever growing amount of textual data and the need to analyze it. One of the important aspects of NLP is computing the similarity between different sentences.
Although it has many applications the different approaches that exists to compare sentences are very naive, most to these methods compute similarity based on common keywords between them. Our approach for comparing the sentences is by considering their underlying context.
The algorithm calculates the similarity score based on not only Syntactic Analysis but by also considering the semantics of the sentence. So the results produced are more accurate as compared to the ones produced just by comparing the Syntax.
We have exposed an API endpoint which you can use to create your own applications using our services.
We have a detailed documentation provided on the webiste which you can refer.
If you just want to try out our services to check first, you can do it in the try it yourself first.
Try it out yourself, register and make request using your own personal API KEY or try the demo directly from the website main page at https://lexicontext.ml