ABSTRACT
With the development of 5G and the emergence of the COVID-19 epidemic, network traffic has surged, and network security has once again become a key concern. Intrusion detection system is an important means to protect network security. It can find abnormal conditions in the early stage of cyber attack. Intrusion detection is also a kind of abnormal detection in a broad sense. To improve the performance of the intrusion detection system, a cyber-attack detection method combining Borderline SMOTE and improved BP neural network (Back-Propagation neural network) is proposed. It mainly uses one-hot encoding, Borderline SMOTE data oversampling and other technologies to preprocess the data, and uses the BP neural network improved by genetic algorithm to predict cyber attacks. Finally, the model is compared with other traditional machine learning models through the core indicator recall and auxiliary indicators precision, roc curve, etc. The results show that the hybrid detection model proposed in this study has higher recall and lower running time, and performs better in intrusion detection. © 2022 IEEE.
ABSTRACT
Drug-target interactions prediction is of great significance in medical and biological research, but traditional laboratory methods have disadvantages such as high cost and time-consuming. Therefore, in recent years, deep learning, similarity calculation methods and other methods are becoming more and more widely applied to related research. This paper proposes an improved deep learning model, named as FPConv-DTI, which uses the fingerprint information of drug and the evolution feature information of protein based on a convolutional neural network. The Borderline-SMOTE algorithm is also used to generate new positive examples for training to solve the imbalance problem, and combines the number of sample data to process the input differently. Experiments have been carried out with four standard datasets and Drugbank dataset. The results show that compared with other methods, our method has greatly improvement for predicting drug-target interactions. In addition, some COVID-19 drugs are also predicted with the best-performing model, which shows that FPConv-DTI model is the potential for practical drug prediction. © 2021 Association for Computing Machinery.
ABSTRACT
Background: The outbreak of COVID-19 attracted the attention of the whole world. Our study aimed to explore the predictors for the survival of patients with COVID-19 by machine learning. Methods: We conducted a retrospective analysis and used the idea of machine learning to train the data of COVID-19 patients in Leishenshan Hospital through the logical regression algorithm provided by scikit-learn. Results: Of 2010 patients, 42 deaths were recorded until March 29, 2020. The mortality rate was 2.09%. There were 6,812 records after data features combination and data arrangement, 3,025 records with high-quality after deleting incomplete data by manual checking, and 5,738 records after data balancing finally by the method of Borderline-1 Smote. The results of 10 times of data training by logistic regression model showed that albumin, saturation of pulse oxygen at admission, alanine aminotransferase, and percentage of neutrophils were possibly associated with the survival of patients. The results of 10 times of data training including age, sex, and height beyond the laboratory measurements showed that percentage of neutrophils, saturation of pulse oxygen at admission, alanine aminotransferase, sex, and albumin were possibly associated with the survival of patients. The rates of precision, recall, and f1-score of the two training models were all higher than 0.9 and relatively stable. Conclusions: We demonstrated that percentage of neutrophils, saturation of pulse oxygen at admission, alanine aminotransferase, sex, and albumin were possibly associated with the survival of patients with COVID-19.