Hot Research Topics in Data Mining and NLP

Natural Language Processing and data mining have been around for a while, and they are both considered as interesting fields to research about. However, it is not easy to find a novel problem or approach for any of them.

In this post we want to talk about some of the “hot topics” in both areas. No matter if you are looking for new research ideas, applications or products that that can help your business or if you are only a data lover, you should keep reading.

  • Chatbots:

Chatbots have become a huge opportunity, and if you are thinking about using them in your daily activity we encourage you to do so! Why are chatbots becoming so relevant? Because it helps companies communicate with their users through new channels as messaging networks. It helps to personalize messages, to provide faster service and to increase loyalty by allowing users to communicate directly with the brand. According to our research and experience conversational UIs have enormous future prospects.

Most of the chatbots available in the market now are based in machine learning and they improve themselves constantly as they keep interacting with more users and learning new words. However, understanding what users are saying is one thing, but knowing what to do is something completely different. There are many ways in which users can ask the same question and the bot may not understand all of them.

The solution and the research area in this topic are related to NLP and also to an applied linguistics approach to enhance user experience while using the bot, so it won't only learn from experience, it will be already taught. In this blog we have talked about chatbots a lot and offered real solutions: fully automated trainingmiddleware to improve chatbot performance, or generation of a training corpus are some of them.


Get to know how to solve 3 of the most common problems in chatbots 


  • Parser development using NLP:

Parsing offers much more possibilities to do text analysis than using any other approach. A clear example of the importance of this tool is that recently Google has created his own parser and also Stanford parser has been around for a while.

What are the interesting topics for research related to parser? Since it’s a tool that is in the growing stage there is still a lot of research to do, particularly related to all the applications parsing may have.

If you want to check out some interesting NLP tools, try our new Framework where you can find: Lemmatization, POS Tagging, Entity Extraction, Phrase Extraction and Sentiment Analysis.



Semantic search is great, but not it isn´t enough anymore. Users should be able to receive condensed information in a structure that is easy to understand, which inevitably requires aggregated facts and knowledge. There has been great progress towards creating greater and greater knowledge databases and interest keeps growing, otherwise Google wouldn't have created its own.

However, there is still a number of open research questions: Knowledge Graphs are often incomplete (so relations and new entities need to be added to the graph, and the graph should be connected to more graphs, i. e., ontology matching), existing information may not be correct (errors may be due, for example, to non-perfect entity resolution, which causes the existence of duplicate entities), and users need to access data in an more accessible way (so it is urgent to work on semantic parsing to better interpret user queries, and on question answering systems built on top of Knowledge Graphs).

A better lemmatization or named entities recognition service would undoubtedly help improvements in this kind of research, and both of them are included in Bitext's NLP Framework. So we invite you to give it a try!



Note: this post was originally published in October, 20th 2016 and has been republished to keep it updated.


Subscribe Here!