Demand prediction model selection method based on demand characteristic analysis

Abstract

The invention discloses a demand prediction model selection method based on demand characteristic analysis, and belongs to the field of demand prediction algorithms. The method comprises the steps: respectively obtaining material data and project data, and carrying out the data cleaning of the material data and the project data; combining the material data and the project data after data cleaning, and obtaining the combined monthly data; inputting the combined monthly data into a prediction model for prediction, and obtaining a weight value corresponding to each prediction algorithm; and determining a final demand prediction model according to the weight values and the prediction algorithms corresponding to the weight values. The historical data is obtained and cleaned, and then is inputted into the prediction model for prediction. The weights of the prediction values and the prediction model for the weight value with the minimum deviation with an actual demand value are selected, and the final demand prediction model is determined. Therefore, the method solves a problem that the actual demand prediction usually has an excessively big prediction error, improves the prediction reliability, is practical simple and is easy to operate.

Description

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine

Login to view more

AI Technical Summary

Benefits of technology

This technology improves how predicts future needs are compared against predetermined goals or standards. It involves gathering relevant past history data and analyzing them beforehand to make informed decisions about whether to use different models depending upon their specific characteristics such as price elasticity (E) versus time-to-demand). Additionally, this innovation allows users to select an optimal model more accurately than previously possible due to its ability to adjust weights according to factors like market conditions without overestimating any expected changes during peak times. Overall, these improvements improve accuracy and efficiency when making accurate predictions while reducing errors caused by unexpected demands.

Problems solved by technology

The technical problem addressed in this patented method relates to accurately estimating how much products will be needed next at any given time without overestimated sales caused by unexpected changes made during previous periods due to poorly chosen prediction models that were used earlier. This can lead to incorrect estimates when selecting an accurate prediction algorithm from limited experimental evidence.

Images