Design And Implementation Of A College Graduate Employment Chatbot For Psychological Counseling

With the social and economic development and the implementation of college enrollment expansion policy,the number of graduates is growing rapidly,the competition in the job market is becoming more and more fiercer,and the psychological pressure of graduates is also increasing.Manual counseling cannot meet the huge demand,so a more efficient and convenient way is needed to deal with it,thus chatbot emerges as The Times require.In natural language processing,the chatbot based on retrieval technology is unable to cope with the scene without pre-defined and has poor flexibility,while the generative chatbot based on deep learning has stronger expansibility and higher system development efficiency,showing a very broad development prospect.To investigate the problems caused by the employment psychological pressure of college graduates in the paper;to use Tensor Flow as the framework for the realization and training of chat module algorithm;to establish a word segmentation dictionary to ensure the correctness of word segmentation;to use stutter segmentation;to adjust the parameters of skip-gram model in word2vec;and to train word vectors.The advantages and disadvantages of generated conversation models such as LSTM model,BILSTM model,BILSTM+Attention model and BILSTM+Attention+Beam Search model were analyzed theoretically.The consistent parameters of each model were adjusted to ensure the influence of the model itself on the training effect,and the above four models were trained on the preprocessed data set respectively.According to the effect of manual evaluation,the optimal model is selected and a comprehensive demand analysis is carried out for the generative chatbot system based on deep learning.Through the demand analysis,the overall framework of the system is proposed and constructed,and each functional module is divided and designed in detail.The front-end framework USES Vue.js and API.Flask framework is adopted for implementation.As for the design of each module,functions of each module are implemented and tested,and the model performance is evaluated by manual evaluation.Tensor Flow framework was used to implement the chatbot conversation model,and the data set in the field of students’ employment psychology was trained.Through the comparative analysis of experimental results,the validity and feasibility of the generative conversation model based on deep learning were verified.Chatbots reduce a lot of labor costs by means of human-computer interaction.