In this study, the classification problem is solved from the view of granular computing. That is, the classification problem is equivalently transformed into the fuzzy granular space to solve. Most classification algorithms are only adopted to handle numerical data; random fuzzy granular decision tree (RFGDT) can handle not only numerical data but also nonnumerical data like information granules. Measures can be taken in four ways as follows. First, an adaptive global random clustering (AGRC) algorithm is proposed, which can adaptively find the optimal cluster centers and maximize the ratio of interclass standard deviation to intraclass standard deviation, and avoid falling into local optimal solution; second, on the basis of AGRC, a parallel model is designed for fuzzy granulation of data to construct granular space, which can greatly enhance the efficiency compared with serial granulation of data; third, in the fuzzy granular space, we design RFGDT to classify the fuzzy granules, which can select important features as tree nodes based on information gain ratio and avoid the problem of overfitting based on the pruning algorithm proposed. Finally, we employ the dataset from UC Irvine Machine Learning Repository for verification. Theory and experimental results prove that RFGDT has high efficiency and accuracy and is robust in solving classification problems.

Pneumonia is an acute respiratory infection that occurs in the alveoli, which is a serious threat to people's health. Nowadays, routine medical diagnosis of pneumonia by taking radiographs of the lungs is an important step, but often due to the interference of other similar underlying diseases or partially inexperienced doctors, etc., in the process of diagnosing pneumonia, which leads to some unavoidable misdiagnosis in the diagnostic process of pneumonia. With the rapid development of artificial intelligence and medical assistant diagnosis, the role of artificial intelligence in the medical field becomes very important. It has great development potential to detect and classify pneumonia images by using the professional and accurate ability of computer. However, due to some defects in the current classic models, such as a small number of AlexNet layers, deleting the convolution layers will seriously affect the classification accuracy, while VggNet, ResNet and other network structure models occupy more resource space due to the increase of convolution layers. Therefore, it is very necessary to design a convolution layer with a moderate number of convolution layers and can better meet the medical needs of pneumonia image classification.

Due to a significant increase in demand for a host of patients for taking lung images and the shortage of the professional imaging doctors in the medical system, it is an urgent problem to be solved that artificial intelligence technology can assist doctors to make a quick judgment on whether a patient is infected with COVD-19 based on medical images. The double-chain network is a new type of network structure, which is composed of two parallel structures for collecting medical images features, supplemented by determining whether it is an infected person. Different from the deeper and deeper training level of artificial intelligence currently, the double-chain structure network is a new type of network that is more efficient, energy-saving and environmentally friendly. The double-chain network structure can play the role of compressing and simplifying the network structure, and the parallel working mode can reduce the redundancy of the network and improve the utilization rate of the network. First, we use a well-known dataset, COVID-CT-Dataset, a CT scan dataset about COVID-19 for medical image classification and recognition. Besides, we use a dataset with uniform positive and negative samples to divide the training set, the validation set and the test set. Finally, a novel double-chain structure network is used to extract features and make classification judgments for CT images.