Financial Crisis Prediction Based on Long-Term and Short-Term Memory Neural Network

Enterprise financial crisis prediction analysis can predict the business process of enterprises, so that enterprises take corresponding strategies in time. The listed companies effectively reflect situation, as to give investors reasonable investment advice. In order supervise sustainable management ability efficiently and accurately, this paper proposed a method based on long-term short-term memory neural network, provide valuable information for decision-makers. Firstly, data enterprise system is analyzed extracted, original cleaned dimensionalized by normalization feature selection. Then, network used build early warning model, wolf pack algorithm optimize initial weight bias parameters, improve efficiency parameter optimization. Finally, 20 large- medium-sized from 2019 2021 are verified analyzed. experimental results show compared with other common machine learning methods, constructed pack-optimized has highest performance terms root mean square error goodness fit, fit reaching 94.2%.