Research on Scientific Typing and Precise Assistance for Small and Micro Enterprises

Small and micro enterprises( SME ) are crucial to the economy and can lead to employment significantly. Their classification provides a comprehensive understanding of the current economic development status, challenges, and opportunities. After indexes system is established, sample classification is obtained by unsupervised cluster algorithm. CatBoost and deep learning like as Long Short-Term Memory(LSTM) and Convolutional Neuron Network(CNN) are compared in prediction performance by confusion matrix. Four stages of SME such as start-up, maintenance-oriented, growth potential, and advanced development types are divided by cluster results and experience. LSTM is preferred to classify SME automatically for improving efficiency. Suggestions and policies about four stages are advised individually.