Working platform task workload prediction method based on deep learning

Abstract

The invention discloses a working platform task workload prediction method based on deep learning. The method comprises the steps of obtaining historical client publishing task data and employee completing task data of a working platform; carrying out missing value interpolation and normalization processing on the client publishing task data; training an LSTM deep learning model as a single-factorprediction model; taking an LSTM deep learning model based on a double-attention mechanism as a multi-factor prediction model; inputting the published task data into a single-factor prediction modelto obtain a prediction result of single-factor prediction; and fusing the prediction result into the lifting tree for regression calculation to obtain a prediction value of the workload. According tothe invention, the single-factor prediction model and the multi-factor prediction model are constructed; a work task failure reason is analyzed, a seq2seq model of the double-attention mechanism is selected, regression calculation is carried out in a lifting tree, a final workload prediction value is obtained, prediction results of multiple models are fused, and the optimal prediction value is solved through cooperation.

Description

technical field [0001] The invention belongs to the technical field of computer software, and in particular relates to a method for predicting the workload of work platform tasks based on deep learning. Background technique [0002] The work platform is an Internet platform that provides various work management related services in a crowdsourcing mode. The contracting party publishes the work task requirements to the work platform, and the platform decomposes the tasks and finds the matching sub-tasks from the platform talent pool according to the skill requirements of each sub-task, and assigns the sub-tasks to the appropriate sub-tasks; The party starts working after receiving the assigned subtasks, and submits the work results to the platform after the subtasks are completed; the contracting party receives and reviews the task delivery results. When the contract issuing party releases the task, it entrusts the task fee on the platform, and after the task is delivered and...

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to provide a method for predicting the workload of work platform tasks based on deep learning. Through the construction of a single-factor prediction model and a multi-factor prediction model, the cause of the failure of the work task is analyzed, and the double attention mechanism is selected. The seq2seq model, regression calculation is performed in the promotion tree, and the prediction value extraction of the final workload of the work platform is obtained, which solves the problems of difficult to predict the workload of the existing customer demand and inaccurate pricing

Images