Research on scientific data mining algorithms based on WOA-BP neural networks

Exploring scientific data is important when dealing with large volumes of scientific data. Scientific data is often difficult to understand, analyse and learn from due to its sheer size and complexity. The aim of this paper is to investigate the WOA-BP algorithm for scientific data mining. The main study is on the regression prediction performance and classification performance of the WOA-BP algorithm on scientific data sets obtained by the WOA intelligent optimisation algorithm after parameter search for BP neural networks. The specific research work of this paper is as follows: (1) Combing relevant research results at home and abroad and summarising existing theoretical foundations. From both domestic and foreign perspectives, the research results on scientific data recommendation, linked data and scientific data management based on linked data are sorted out, and the already developed and mature theoretical foundations of scientific data, scientific data mining, linked data and related concepts are summarised. (2) Construct a scientific data mining model. Firstly, the principles of model design are clarified, and then the general architecture of the scientific data mining model based on open source scientific datasets is constructed, which is divided into BP neural network model and WOA intelligent optimization model. Several groups of benchmark control models were selected for comparison experiments, and the final WOA-BP algorithm regression superiority R2 value reached 0.944 and classification superiority accuracy value reached 93.691%, both of which were higher than other benchmark comparison models, fully proving the superiority of the WOA-BP model proposed in this paper.