A Robust Evolutionary Gravity Neocognitron Neural Network Model for Construction Cash Flow Prediction in Complex Project Management

In order to keep expenses in check while building projects are being carried out, accurate cash flow forecasting is crucial. The project's unique features and risk considerations make it challenging, despite extensive study, to develop a reliable model for predicting future cash flows. Furthermore, both sequence and non-sequence elements have a significant influence on the building projects' cash flow. As a result, the authors of this research suggested a new inference model based on artificial intelligence (AI) called Evolutionary Gravity Neocognitron Neural Network (EG-NN). This model takes project complexity into account and attempts to handle non-sequence elements in time series issues. By attaining the highest values of 2.35%, MAPE of 5.65%, MAE of 2.03%, and coefficient of determination of 0.99, the experimental findings on thirteen construction projects have validated the EG-NN as the top model for cash flow forecasting. When compared to other AI-based inference models, the statistical result shows that cash flow forecasting can be improved by at least 14.1 percent in terms of root-mean-square error (RMSE) and 15.2 percent in terms of mean absolute error (MAE). Managers may benefit from the EG-NN network model's ability to predict and manage construction projects' cash flow.