Regular Expressions For Improving Natural Language Processing

   Source: http://www.expertsystem.com/natural-language-processing/

Natural Language Understanding (NLP) comes under the category of Computer Science and Artificial Intelligence. First, let us understand what exactly is NLP.  In simple words, NLP means processing natural language to extract relevant information. It is concerned with communication between human and computer. NLP allows humans to talk to machines i.e. it makes a machine capable of understanding human language as it is spoken. NLP performs analysis and synthesis on natural language and speech using some computational techniques. 

But how do NLP works and how Natural Language is converted into Text for processing:

1-   First step is to extract some meaning from natural language that the computer received. For this purpose, Automatic Speech Recognition(ASR) or Speech to Text is used, which takes natural language and convert it into artificial language i.e. converts the information into text format.

2-   Once we have information in text, NLP uses different techniques to understand the text which is generally a request from a user and generate a response corresponding to the request which is coded in the system.

Since there are many techniques for understanding text and generating a response. Here we will be talking about how can we improve the performance of NLP for response generation which is also called Natural Language Generation(NLG). For that searching technique should be efficient enough to extract relevant response. 

We will be here discussing how Regular Expression(Regex) can be used for improving searching technique and hence NLG. Regex is basically a string of text allowing a user to create a pattern for matching, locating and managing text.

Since using  ASR, we can have our information in the form of text. Text can be viewed as a string of characters. Regex here is incredibly useful because it provides us numerous pattern for matching. Regex is important because searches can have many forms. Some of the examples are given below.

We have some information corresponding to text:- How's the weather outside?
 

Now user can search this information in any form:- How is the weather outside? or only Weather Outside or only Weather. Regex patterns help us to reduce stemming and match all type of searches which mean the same. 

This is how Regex can be used for improving NLP.