Research On The Push Algorithm Of Take-out Rider Order

With the emergence and popularity of mobile Internet technology,take-away consumption has become a new fashion.Because there are many objects involved in the field of take-out,there are businesses,riders and users,and it has a complex relationship network,efficient delivery experience,and high-quality user experience.The dynamic adjustment of the merchant's distribution range enables the order structure to be optimized;the order's operational algorithm,the rider's path planning algorithm and iterative strategy,and enhanced learning techniques also greatly improve the rider's distribution efficiency.For a fast delivery order,the difference in order duration is a reflection of the rider's preference for the order.Therefore,this paper hopes to dig out the relationship between the order time of the rider and the order from a large number of historical express delivery orders,and dig out the characteristics of the orders with shorter time for the rider to take orders,and analyze these characteristics with the time of the rider.The relationship,in turn,predicts the order length of the order by the order characteristics,so as to reduce the proportion of orders for the rider to take longer orders in the take-out order,and increase the proportion of orders with shorter orders for the rider in the take-out order.In turn,the distribution efficiency of the overall take-out order is improved,and the user experience is improved.The algorithm consists of two layers of architecture.The first layer uses several conventional classification algorithms.For each classification algorithm,the training set is cross-trained,and each time the test result is used as a feature in the new training set;then the trained classification algorithm is used.Test and use the predictions as a feature in the new test set.For the second layer classification algorithm,the new training set training algorithm obtained by the first layer classification algorithm is applied,and the new test set is used to verify the algorithm effect.The main function of the paper is based on a large number of historical take-out order data,from the characteristics of each dimension of the order to explore the inherent relationship between the rider's order time and the order,through the machine learning andother technical means to identify the user's order of the rider orders The duration is designed to reduce the proportion of orders for the rider to take longer orders,and to increase the proportion of orders for which the rider receives shorter orders,thereby improving the efficiency of the rider and further improving the user experience.