Building NLP model for classifying short-tail conversational student’s query data

Kural Arasu Venkatesh, B.E.

Master's thesis

Abstract:

The objective of this thesis is to construct a Natural Language Processing (NLP) model capable of effectively categorizing students' queries based on their respective intents, thereby enabling their efficient and accurate routing to the appropriate department. To accomplish this, an LSTM model was selected owing to its well-established capacity for handling textual data. Prior to feeding the data into the neural network, pre-processing was carried out to convert the inputs into numerical values that could be processed by the neurons. The training and evaluation of the model were performed using two distinct datasets, with Accuracy being the primary performance metric. Additionally, a comparative analysis was conducted by comparing the outcomes of the LSTM model with those of a pre-existing model in the market, IBM Watson. The results demonstrated that the LSTM model outperformed IBM Watson in accurately classifying students' queries into their respective intents. The proposed classification model has considerable potential for use in university chatbots, as it could substantially aid prospective students and academic institutions in streamlining query organization and improving overall operational efficiency.

Keywords

Natural Language Processing, Artificial Intelligence, Long-Short- Term-Memory

Language used: English

Date on which the thesis was submitted / produced: 3. 4. 2023

Thesis defence

• Date of defence: 24. 5. 2023
• Supervisor: doc. Ing. Jan Tyrychtr, Ph.D.
• Reader: Richard Vágner, externi

Czech University of Life Sciences Prague

Faculty of Economics and Management

Master programme:
Informatics