Profit prediction optimization using financial accounting information system by optimized DLSTM.

Area of Science:

• Financial Accounting Information Systems (FAISs)
• Computational Intelligence
• Machine Learning

Background:

• Deep learning (DL) and swarm-based algorithms are increasingly used in FAISs.
• High complexity of extensive datasets poses challenges for hybrid networks in financial forecasting.

Purpose of the Study:

• To develop an advanced methodology for financial accounting profit prediction.
• To enhance the efficacy of profit prediction models using hybrid DL and optimization algorithms.

Main Methods:

• Integration of the twin adjustable reinforced chimp optimization algorithm (TAR-CHOA) with deep long short-term memory (DLSTM).
• Development of a new dataset using fifteen inputs from the Chinese stock market Kaggle dataset.
• Design and assessment of five DLSTM-based optimization algorithms for profit forecasting.

Main Results:

• The DLSTM-TAR-CHOA model achieved the highest performance among the evaluated DL-based models for financial accounting profit prediction.
• The newly developed TAR-CHOA algorithm significantly improved the efficacy of profit prediction models.

Conclusions:

• The DLSTM-TAR-CHOA model represents a state-of-the-art approach for financial accounting profit prediction.
• Future research should explore alternative methodologies to address data dependency and market variations.