A quantitative structure-property relationship (QSPR) model is proposed to explore the relationship between the pKa of various compounds and their structures. Through QSPR studies, the relationship between the structure and properties can be obtained. In this study, a novel chaos-enhanced accelerated particle swarm algorithm (CAPSO) is adopted to screen molecular descriptors and optimize the weights of back propagation artificial neural network (BP ANN). Then, the QSPR model based on CAPSO and BP ANN is proposed and named the CAPSO BP ANN model. The prediction experiment showed that the CAPSO algorithm was a reliable method for screening molecular descriptors. The five molecular descriptors obtained by the CAPSO algorithm could well characterize the molecular structure of each compound in pKa prediction. The experimental results also showed that the CAPSO BP ANN model exhibited good performance in predicting the pKa values of various compounds. The absolute mean relative error, root mean square error, and square correlation coefficient are respectively 0.5364, 0.0632, and 0.9438, indicating the high prediction accuracy. The proposed hybrid intelligent model can be applied in engineering design and the prediction of physical and chemical properties.

Solubility of gas in polymers is an important physico-chemical property of foam materials and widely used in the preparation and modification of new materials. Under the conditions of high temperature and high pressure, the dissolution process is a nonlinear, non-equilibrium and dynamic process, so it is difficult to establish an accurate solubility calculation model. Inspired by particle dynamics and evolutionary algorithm, this paper proposes a hybrid model based on chaotic self-adaptive particle dynamics evolutionary algorithm (CSA-PD-EA), which can use the iterative process of particles in evolutionary algorithms at the dynamic level to simulate the mutual diffusion process of molecules during dissolution. The predicted solubility of supercritical CO ₂ in poly(d,l-lactide- co -glycolide), poly(l-lactide) and poly(vinyl acetate) indicated that the comprehensive prediction performance of the CSA-PD-EA model was high. The calculation error and correlation coefficient were, respectively, 0.3842 and 0.9187. The CSA-PD-EA model showed prominent advantages in accuracy, efficiency and correlation over other computational models, and its calculation time was 4.144–15.012% of that of other dynamic models. The CSA-PD-EA model has wide application prospects in the computation of physical and chemical properties and can provide the basis for the theoretical calculation of multi-scale complex systems in chemistry, materials, biology and physics.

Biological sequence data mining is a long-term hot spot in bioinformation. Biological sequence can be regarded as a set of characters composed of a number of letters and contain an evolutionary relationship. Time series is a set of numbers arranged according to time and contains the temporal progressive relationship. Time series is similar to biological sequence in terms of both representation and mechanism. Therefore, in the paper, biological sequence is represented with time series to form biological time sequence (BTS). Based on advanced time series methods, hybrid ensemble learning framework (SaPt-CNN-LSTM-AR-EA) for BTS is proposed. Single-sequence and multi-sequence models are constructed based on self-adaption pre-training one-dimensional convolutional recurrent neural network (CNN-LSTM) and autoregressive fractional integrated moving average (ARFIMA) integrated evolutionary algorithm, respectively. In the DNA sequence experiments of six kinds of viruses, SaPt-CNN-LSTM-AR-EA realized the good overall prediction performance, the prediction accuracy and correlation were 1.7073 and 0.9186, respectively. The effectiveness and stability of SaPt-CNN-LSTM-AR-EA were verified through the comparison with other five benchmark models. In addition, compared with other benchmark models, SaPt-CNN-LSTM-AR-EA increased the average accuracy by about 30%. This study opened up a new field of BTS research. The framework proposed in this paper is significant in many disciplines, such as biology, biomedicine, computer science and economics. Especially in sequence splicing, genome, computational biology, bioinformation, theoretical biology, evolutionary biology, signal processing, medicine and health care and other fields, the framework has a wide range of applications.

Quantitative structure-activity relationship (QSAR) model is adopted to study the relationship between the chemical and physical properties of various substances and the structure. Through QSAR studies, the internal relationship between the invisible structure and the activity can be obtained. In this paper, a novel chaos-enhanced accelerate particle swarm algorithm (CAPSO) is proposed, which is used to molecular descriptors screening and optimization of the weights of back propagation artificial neural network (BP ANN). Then, the QSAR model based on CAPSO and BP ANN is put forward, hereinafter referred to as CAPSO BP ANN model. The prediction experiment showed that the CAPSO algorithm is a reliable method for screening molecular descriptors and the five molecular descriptors obtained by CAPSO algorithm could well characterize the molecular structure of each compound in pKa prediction. The experimental results also showed the CAPSO BP ANN model has a good performance in predicting the pKa values of various compounds, the absolute mean relative error, root mean square error, and square correlation coefficient are respectively 0.5364, 0.0632, and 0.9438, indicating the higher prediction accuracy and correlation. The proposed hybrid intelligent model can be applied in all kinds of engineering design, prediction of physical and chemical properties and intelligent calculation.