Deep Learning Based Natural Language Processing E-Commerce Chatbot

Title

Deep Learning Based Natural Language Processing E-Commerce Chatbot

Authors

1. Rudri Jani, Ahmedabad University, Developer, India
2. Nikit Patel, Cygnet Infotech, Developer, India

Abstract

Nowadays, chatbots are becoming popular because it feels like talking with

humans in natural language in live chat. This research aims to develop an

eCommerce chatbot to help customers buy different pet products from online

pet stores. The system uses deep learning and natural language processing

concepts like text classification, intent classification, multiclass classification,

named entity recognition, etc. The system will guide the customer through

purchasing the products by asking questions and replying. The system will also

extract the entities from the user queries through NER and display the products

accordingly to the user’s request. It will interestingly engage the user in buying

pet products by solving the user’s query and providing a quick response to the

user. This will also increase sales, raising the revenue of the company.

Keywords

chatbot deep learning natural language processing named en- tity recognition eCommerce text classification multiclass classification rule- based chatbot ElasticSearch MongoDB

PDF

This browser does not support PDFs. Please download the PDF to view it: View the PDF.

Conclusion

Creating the hybrid model for the chatbot using rule-based and AI-based ap-

proaches rather than only making it rule-based or AI-based has its own ad-

vantages. Using a self-learning approach, a chatbot can reply to the user by

identifying the intent of the trained model. Using a rule-based approach, the

task is performed accordingly if certain conditions are matched. Therefore com-

bining two models and using a hybrid system to develop the chatbot has an

advantage. Using the NER model helps to filter out the entities and search the

products from the database. Introducing the elasticsearch helps to improve the

search efficiency instead of just using regular expressions to find the products

from the database.

Reference

1. [1] A. Roy, “Designing a chatbot using python: A modified approach,” Jul 2020. [2] M. Grunitz, “Rule-based AI vs machine learning: what’s the dif- ference?.” https://wearebrain.com/blog/ai-data-science/rule-based-ai-vs-machine-learning-whats-the-difference/, Sept. 2021. Accessed: 2022-4-26. [3] R. Agrawal, “Must known techniques for text preprocessing in NLP.” https://www.analyticsvidhya.com/blog/2021/06/must-known-techniques-for-text-preprocessing-in-nlp/, June 2021. Accessed: 2022-4-26. [4] Harshith, “Text preprocessing in natural language processing using python,” May 2021. [5] dishashree26, “25 must know concepts for beginners in deep learning amp; neural network,” Jun 2019. [6] “Data formats · spacy API documentation.” https://spacy.io/api/data-formats. Accessed: 2022-4-26. [7] Shrivarsheni, “Training custom ner models in spacy to auto-detect named entities [complete guide],” Apr 2022. [8] “What is mongodb?.” [9] “Artificial linguistic internet computer entity,” Dec 2021. [10] “Difference between IOB and IOB2 format?.” https://datascience.stackexchange.com/questions/37824/difference-between-iob-and-iob2-format. Accessed: 2022-4-26. [11] “What is elasticsearch?.” [12] “Dialogflow documentation nbsp;—nbsp; google cloud.” [13] “Tokenizer reference: Elasticsearch guide [8.1].”

Author Contribution

Rudri Jani has done all research and has participated in the conception and execution of study. Nikit Patel has managed and guided her in the research.

Funding

No funding was provided to the author(s) of this article during its research, writing, or publishing.

Software Information

Conflict of Interest

Each author confirms that they have no competing interests.

Acknowledge

My mentor Nikit Patel, has been a constant source of inspiration, encouragement, advice, excitement, and belief in my abilities. His knowledge in the IT business, ideas, and suggestions on introducing new things were quite beneficial to me. I’d want to thank him for devoting some of his valuable time to interns like us.

Data availability

Data was created by Rudri Jani relevant to the models trained for the implementation. But the data is confidential so it can't be shared.