Research On Early Warning Method Of Railway Bulk Cargo Freight Price Risk Based On Deep Learning

Bulk goods include coal,steel,crude oil,etc.,which are important raw materials that affect the economy and people’s livelihood.Railway transportation has the characteristics of large transportation capacity,low transportation cost,and high safety.It is the main mode of transportation of bulk goods in our country and the main artery of the national economy.In recent years,with the overall advancement of market-oriented reforms,competition among various modes of transportation has become increasingly fierce,and railway enterprises have also been greatly challenged.In order to expand market share and increase corporate profits,formulating reasonable freight rates is an extremely critical part.Although the Chinese government has relaxed its control over railway enterprises and allowed them to have independent pricing power within a certain floating range,railway transportation is still at a disadvantage in the freight market competition.To this end,this paper analyzes the current situation of my country’s freight market and draws on relevant research experience at home and abroad,and proposes to establish a railway freight rate early warning mechanism to assist the railway sector in formulating reasonable freight rate adjustment strategies to avoid the occurrence of freight rate risks.First,this article analyzes the changes in railway transportation in the freight market in recent years,and analyzes the problems in the pricing of railway bulk freight and the necessity of early warning.It analyzes the impact of the macro economy,the transportation market,the cargo owners and the railway enterprises on the price of railway freight,and builds the early warning index system for the price of railway bulk freight based on this.Combining the entropy weight method to calculate the weight of each indicator,and propose a risk alert division method based on the K-means clustering method.Secondly,by studying the related theories of deep learning,this paper proposes three corresponding improvement methods for the gradient problem and over-fitting phenomenon in the training process of LSTM neural network.This paper also established a railway bulk freight price risk early warning model based on BP neural network and improved LSTM neural network,and made it into memory through network training.In this way,the model can output the freight rate risk alert for the next month by reading the index data of the current month,so as to achieve the effect of early warning.Finally,this article uses the coal transportation situation of S Railway Bureau in2015-2017 as an example to verify the feasibility of the model established in the previous article.In this paper,70% of the data samples are selected as the training of the network,and 30% of the data samples are used as the test.The results show that the performance of the optimized LSTM neural network has significant advantages compared with the optimized LSTM neural network and the BP neural network.It can more accurately predict the next stage of freight risk level based on the current indicator status,thereby assisting relevant departments Work out a more reasonable tariff adjustment strategy to scientifically resolve the risk of freight rates.