Abstract
A chatbot is a computer program that interacts with users in natural language under a simple interface such as audio or in the form of a message. Chatbots are often used in voice guardian systems for various practical purposes including customer services, consulting services, and information collection. The Chatbot system uses natural language processing, the system simply scans the input keywords by entering text directly and then the system will select an answer with the most suitable keywords or the most similar word pattern in the training database set to answer the user.The career counseling system for high school students written on the Google Colab platform is a free cloud service, supported by GPU and TPU. Google Colab comes pre-installed with many popular Deep Learning libraries such as PyTorch, TensorFlow, Keras, and OpenCV. This study uses libraries used for NLP natural language processing, TensorFlow library, Adam optimization algorithm, network_DNN neural network, and AI Chatbot automated technology to perform interactive career counseling between users and the system based on data sets collected from several industry groups.
Introduction
Career counseling and support for high school students are crucial for helping them transition to higher education or vocational training that aligns with their interests and personalities. However, traditional methods, such as organizing visits to colleges and universities, are often costly and inefficient, lacking in providing comprehensive information due to constraints like time and participation conditions. Nguyen Thong High School has a total of about 1400 students, and each year it recruits about 450 new students according to the target of the Department of Education and Training of Vinh Long province. However, every year, there are many senior students who cannot determine the profession they need to study; a few still choose at the request of their parents or choose a profession according to the majority. Although school management and homeroom teachers always focus on career counseling, the counseling work is not really effective because it depends on the learning ability and interests of students. The school also does not have a department of career counseling experts for students, mainly organizing collective career counseling sessions or organizing admissions counseling tours organized by universities to advise the school on admissions. That has not made them understand the professions they want to study, have not absorbed information, and are not clearly aware, leading to not being able to choose the right profession that they can learn and be useful for themselves.
Proposed Method
The main function of the system is to receive messages, process them, and respond to user requests. Initially designed to meet the basic requirements of a career counseling system for high school students, the chatbot focuses on a few key areas: guiding students in selecting careers that match their abilities and strengths, advising on suitable subject combinations for university admissions, and providing relevant information about three major industry groups—engineering, information technology, and the social sector.
The web-based chatbot system has a simple interface that supports communication with users through a number of consulting questions in 3 groups of industries. The conversation process is carried out by entering Vietnamese text with accents, and the results of answering questions are displayed as text for users to read on their own. In order to have a truly intelligent chatbot that can replace humans conducting conversations for career counseling, it is necessary to have a rich knowledge base so that conversations will always be continued according to a certain topic that both parties are interested in. Initially, creating an effective career counseling "robot" for high school students. Based on the analysis of chatbots in the above section, the thesis chooses a solution that combines open source: using open-source libraries on artificial intelligence and deep learning to support programming to solve problems: classifying user intent, extracting information, and calculating the similarity of texts.
Results
The study conducted an experimental run on 310 sample sentences, with the results evaluated as shown in Table 1.
Through testing on different epochs, it was found that at the last run, the epoch was equal to 2000, corresponding to the number of training steps at 8000. The result achieved an accuracy of up to 99.27%, of which the lowest loss index was estimated at 0.3912%.
Conclusions
Vocational education is crucial in shaping a country's workforce, and vocational counseling has emerged as an effective modern approach utilized by many countries worldwide. However, in our country, the use of vocational counseling, especially for high school students, remains limited and underexplored. Through theoretical research on vocational education, I have identified vocational counseling, supported by counseling system software, as an independent path within career education. This approach has distinct goals and content, following a structured process that maximizes its benefits and meets vocational education objectives in schools. In my research, I developed a “Career Counseling Support System for High School Students” using chatbot technology, yielding promising results. The thesis delved into natural language processing, text classification techniques, and chatbot building methodologies, proposing a model suitable for career counseling. With a dataset of over 310 sentences across various topics, the system successfully provided guidance in three major industry groups: engineering, information technology, and social sciences, achieving an accuracy of 99.27% when user inquiries matched the database's scope. The dataset's flexibility allows for the addition of more topics, enhancing the system's accuracy. Consequently, chatbot technology can offer students valuable guidance in choosing suitable careers after high school.
References
- Bhagwat, Vyas Ajay, "Deep Learning for Chatbots" (2018). Master's Projects.
- ChatBots ORG Homapge. https://www.chatbots.org/ retrieval.
- Mikhail Burtsev, Alexander Seliverstov, Rafael Airapetyan, Mikhail Arkhipov, Dilyara Baymurzina, Nickolay Bushkov, Olga Gureenkova, Taras Khakhulin, Yuri Kuratov, Denis Kuznetsov, Alexey Litinsky, Varvara Logacheva, Alexey Lymar, Valentin Malykh, Maxim Petrov, Vadim Polulyakh, Leonid Pugachev, Alexey Sorokin, Maria Vikhreva and Marat Zaynutdinov, June 15, 2018: "DeepPavlov: Open-Source Library for Dialogue Systems”
- S. Fabrizio (2002), Machine Learning in Automated Text Categorization.
- C.D. Manning, P. Raghavan and H. Schütze (2008), Introduction to Information RetrievalIntroduction.
- D. Jurafsky, J-H. Martin (2017), Speech and Language Processing. Computer Science, Stanford University