Unicorn NLP API converts any unstructured Review (or any other text) into structured ready-to-use data which could be easily queried, analyzed and displayed - in the same way like normal structured data. It is the next-generation of Sentiment Analysis.
It could be used for enhancing user-interface, analytics, next-gen reviews presentation for mobile, review summary, additional filters for browsing or semantic search based on reviews and more.
The API (accessible via RapidAPI) is available for Hotel Reviews/Opinions and we are working on other domains. If you need to process a different kind of texts, we can prepare custom made NLP engine tailored for Your requirements for any domain, any kind of texts.
We have achieved state-of-the-art accuracy for Hotel Reviews which is far better than any tool available.
Precision=95%, Recall=80%, F1=87%
(If we extract 100 results, 95 of them are correct)
We have 10+ years expierience as NLP engineers and we had to use all of our knowledge to get to this point. We have built the engine from scratch and we use a totally different approach than mainstream or what is described in scientific publications. And we have succeded. All we can say is that we have used a lot of handcrafted semantic rules, which are designed and widely tested in specific domain.
If You do not have NLP team in your company or need to outsource something this is the perfect solution. We are able to build engine for any kind of texts for English language. You deliver the data, specify which information you want to extract and we prepare the engine with very high accuracy. You could also send us only the data and ask us to analyze and specify what kind of information we think it would be the best to extract. Typically it takes 3 months, but first working version You get after 1 month to make sure that we going in the right direction. We deliver 2nd version after 60 days and final version in 90 days. Custom-made NLP libraries start from 25k USD.
We do not collect your data while we process, our goal is to make the best NLP engine and leave the business to You. We analyze statistical data and check for errors (and correct them if we found them) to ensure that the high quality is maintained. If Your data is specific and cannot be send outside the company, we can deliver the engine, which will be hosted on Your servers. Your choice.
Instead of mathematics description used in Sentiment Analysis i.e. Breakfast(-2,-1,0,1,2) we believe in a non-linear human-based description like Breakfast (great, rich, small portions, not provided, extra paid etc.) or Room (great view, clean, spacious, noisy, renovated etc.) or Opinion (will return, will recommend, overal good/satisfied, will not return, etc.) which is more relevant description of the human world. It is compatible with Aspect-based Sentiment Analysis but also provide you a deeper human-like level of understanding. Output is accurate and simple enough to display it to the user on mobile.
Ready to use output from our API does not require linguistic knowledge or data science knowledge to understand or to integrate API into Your application. Our output is simple and intuituve and you do not have to know anything about part-of-speech or machine learning or other sophisticated knowledge. We spent a lot of time to hide all the complexity inside so the output only gives ready-to-use actionable data. If you have linguistic knowledge, great. Our output could be easily connected with Your NLP data and it should provide a lot of new options for Your application.
The API for Hotel reviews (Human-like Sentiment Analysis for Hotel Reviews API) is accessible via Rapid API and you pay only for processed Reviews. It is possible to buy the on-premise version of the engine (PHP 7), which can be used without limitations. You can install it wherever you want and process whatever you want without sending to our servers. If you are interested in the second solution, write to us: email@example.com
Comparison of our technology vs. Sentiment Analysis - read more...
Go to RapidAPI and start using this API today.