Method and device used for carrying out prediction on predicted object

Abstract

The invention provides a method and a device for carrying out prediction on a predicted object. The method comprises the steps of acquiring a plurality of historical data values of the predicted object in a predetermined historical period; searching a plurality of influence factors related with the predicted object from a database, and acquiring one or more data sources corresponding to each influence factor and a predetermined algorithm; searching text information of each influence factor from the one or more data sources, and converting the text information into a plurality of pieces of structured data according to the predetermined algorithm; acquiring a correlation coefficient of each influence factor and the predicted object according to the plurality of historical data values and the plurality of pieces of structured data, and taking out part of the influence factors whose correlation coefficients meet a predetermined condition from the plurality of influence factors; acquiring the length of a prediction period which is immediately after the predetermined historical period according to the plurality of historical data values, and calculating the weights of part of the influence factors respectively; and acquiring a plurality of prediction data values of the predicted object in the prediction period according to the length of the prediction period and the respective weights.

Description

technical field [0001] The present invention relates to a method and device for forecasting general demand, in particular to a method and device for forecasting objects. Background technique [0002] With the continuous expansion of the scale of modern production and sales, how to make a reasonable production plan has become a major issue that affects the cost and profit of the enterprise. Many companies do not have an effective method to predict future market demand, or the backlog of excess products in the warehouse has brought great pressure to the capital chain, or the supply of products must not miss the effective opportunity to occupy the market. These possible results will seriously affect the operating efficiency of the enterprise and affect the long-term development of the enterprise. [0003] As an important part of the field of artificial intelligence, demand forecasting technology has been studied for many years, and many related technologies have been proposed....

AI Technical Summary

Problems solved by technology

[0004] However, most of the existing forecasting technologies face the following problems: 1. The forecasting technology is not universal enough, and most forecasting tools are designed for a specific field to select forecasting impact factors and design forecasting models

The prediction tools developed in this way are difficult to apply to other fields, and repeated development is also a waste of resources

2. Relying too much on data integrity when making forecasts

In many practical applications, customers cannot provide the data required for complete prediction, so it is impossible to train a high-precision prediction model

Images