Research On Topic Classification In The Field Of Renewable Energy Based On LSTM And LDA

Nowadays,the rapid growth of scientific research literature in various disciplines and the deepening intersection between disciplines show that its branches are numerous and development is uneven.This situation makes it more difficult for scientific researchers to accurately search for information,and it is not conducive to clarifying the research progress of disciplines and tracking the frontiers in the field.This thesis takes the field of renewable energy in the electrical and electronics discipline as an example,and applies the Latent Dirichlet Allocation(LDA)and Long Short-Term Memory(LSTM)methods to the topic distribution detection in this field.As a new attempt in this important field,aimed to provide academic researchers with the clearest development context and organizational structure in this field.First,we applied statistical analysis,latent dirichlet allocation(LDA)topic modeling technique,and autoregressive integrated moving average(ARIMA)model to map topic landscapes in the field of renewable energy.By analyzing,we discovered the digital characteristics of the field and divided the field into 29 different topics and analyzed the growth characteristics of topics in two time periods.Finally,based on the development trajectory of each topic,we predicted their future development enthusiasm,which was divided into cold,hot and stable.We compiled statistics on the most popular outlets and citations in each topic,making it easy for researchers and journal editors to appreciate and apply.This thesis analyzes the most discussed topics,hot topics and cold topics based on the above calculation examples combined with existing technology and literature,and prospects the technology development in the research field.Secondly,this thesis uses the word2vec+LSTM model to conduct another classified study on the field of renewable energy in the electrical and electronic disciplines.By manually labeling the training data and mapping the word vectors,and then classifying the research field based on the LSTM model,the literature in the field of renewable energy in the electrical and electronic disciplines is divided into 17 different topics,and then uses LSTM to predict the academic research trends in this field.Finally,this article compares LDA and LSTM’s topic classification work in this field from multiple angles,and the results can be used as a reference for all subject areas.On this basis,based on the rapid growth of the subject literature collection and the complexity of the topics included,this thesis proposes an LDA and LSTM fusion topic classification model suitable for subject topic classification studies.In addition,based on the fusion model,the data of the research field in 2019 was classified.