Research And Implementation Of Timing Analysis Based On Large Scale Express Data

According to the State Post Bureau data show that China's express delivery industry is developing rapidly,and has maintained an explosive growth of about 50%for 6 consecutive years,injecting a strong vitality into the economic growth of our country.However,with the rapid development the more and more problems have come,such as express warehouse explosion,random stacking,violent sorting,delayed delivery,etc.,resulting in many courier was destroyed,wrong delivery,delay.Especially when it is not the usual date,such as double eleven,before and after the Spring Festival and so on,due to the courier traffic,logistics personnel,sorting staff,dispatch staff,storage capacity,etc.are not allowed to estimate,often resulting in poor service consumer experience.In order to better predict the various resources needed by the express delivery industry,the most fundamental purpose is to accurately predict the express delivery volume and provide decision support for optimizing resource allocation.In order to solve the above problem,this paper proposes a multi-feature LSTM-based courier traffic sequence prediction model for the first time,constructing the festival features according to the characteristics of each festival,and compares it with the time series prediction model based on the featureless LSTM and ARIMA to make comparative experiments at different time points,such as regular day,Spring Festival,National Day,and National Day.At the same time,this paper also established a gray correlation analysis model and use this model to further explore the impact of express delivery business.The experimental verification shows that LSTM-based express traffic sequence prediction model is far superior to ARIMA-based express traffic sequence prediction model under MSE,RMSE,MAE and R-Square four indicators.And after adding the relevant festival features,the LSTM model can also more finely depict the business trends of the Spring Festival,the National Day and the double-11.At the same time,based on the gray relational degree model to analyze the influencing factors of express traffic,the importance of influencing factors of express traffic is in turn Internet,economy,infrastructure.To facilitate the using and analysing of express traffic data for State Post Bureau staff,this paper also developed LSTM-based express traffic prediction function module,and integrated in the courier data offline analysis system.