Comparative Study on Cold Chain Logistics Based on Neural Network Word Embedding

• 1、School of Management, Wuhan University of Science and Technology, Wuhan, 430065, China

Abstract：With the rapid development of e-commerce and global trade, cold chain logistics, as an important technical means to ensure that perishable goods such as fresh food and drugs maintain a specific temperature during transportation and storage, has gradually become a key link in supply chain management, whose efficiency and reliability are directly related to consumer satisfaction and enterprise economic benefits. This paper uses neural network word embedding Word2Vec method to compare academic platform and media platform around cold chain logistics, aiming to promote the efficient development of cold chain logistics. First of all, we choose CNKI and China Logistics and Procurement Network as comparative research objects, and combine these two platforms to collect data around cold chain logistics and form an academic corpus and a media corpus. Then, Word2Vec modeling is carried out on these two corpora respectively. On this basis, combining the top 20 words that are most similar to "logistics" and the top 20 words that are most similar to "cold chain logistics", the focus of the two platforms is compared. Using t-SNE algorithm to reduce the dimension of the word vector, using matplotlib.pyplot package to visually display and compare the word embedding profiles of all the words in the two Word2Vec models. The results show that cold chain logistics is different between academic platform and media platform. This research innovatively uses Word2Vec algorithm to carry out comparative research, which provides a new perspective for cold chain logistics research.

Keywords： Cold chain logistics Word2Vec Comparative study Academic platform Media platform