Forecasting coordinated actions of actors in a software product with a network effect

The article explores the challenge of increasing actor activity in software products with network effects, followed by the spread of their actions beyond these products, based on the application of a modified forecasting method. Activity in software products with network effects is considered in the context of complex nonlinear dynamic systems. To forecast a sharp increase in actor activity, an existing method for predicting abrupt changes in time series has been modified. The proposed method was tested using open data on comments from the wallstreetbets community on the Reddit social network. Additionally, data on the stock prices of GameStop Corp (GME) for the year 2020 and the first two months of 2021 were utilized. The results of the testing confirmed the effectiveness of the proposed approach in addressing specific applied problems, particularly those related to the speed of response to the development of “negative” events, socially significant topics, and the prevention of offenses and social unrest.