This repository is dedicated for chat-bot university thesis project for earning Data Analytics master degree
Users are rapidly turning to company’s customer care to receive customer support about their difficulties while using any products or services. However, most of these queries are not being addressed on time or in some cases not addressed at all. Which leads to customer dis-satisfaction. COMPRA GmbH is one of those companies, which is also witnessing this problem. With the increase in number of user base and services, the customer support which is being done via phone, emails and ticket systems gets unnecessarily time-consuming. Another problem is to have human assistants in customer support and the cost it causes to the company. One way to solve this problem is to create a new conversational system to automatically generate responses for customers requests via a chat-Bot. Currently, in domain specific conversational systems structured data is being used for the bot to train however, gathering such a data-set is not an easy task for many companies. Preparing training data for chat-bots in conventional format itself is a time consuming and costly process. This thesis present a solution for data structurization for chatbot training where bot will be able to understand user inputs and answer them accordingly. The goal is to develop a chatbot from unstructured data by building a novel framework for feature extraction from such data. First, the performance will be tested by human analyst within the COMPRA GmbH and then we will deploy the COMPRA bot into real world an evaluate it’s performance on genuine customers’ queries.
Since there are no such algorithm exist, I am trying to combine few algorithm to make it happen. Currently, almost all of the chatbot works with structured information. Machine Comprehension can be one thing to work with, in this situation. Stay tuned for future updates.
+ Please give me some idea if you have read the proposal. I would highly appreciate that.