Construction and Quality Inspection of an Optimization Model for the Survival Conditions of Animal like Earthworm Traditional Chinese Medicine Based on BP Neural Network

As a traditional animal based traditional Chinese medicinal herb, the living conditions of earthworms (temperature, humidity, soil pH, etc.) directly affect their survival rate, reproduction rate, enzyme activity content, and medicinal quality. In response to the current lack of standardization and significant quality fluctuations in aquaculture conditions, this study proposes a survival condition optimization model based on BP neural network and a multi index quality detection method. By conducting a multi factor experimental design to collect data on earthworm farming, a BP neural network model was constructed with environmental parameters (temperature, humidity, pH, feed ratio) as input layers and mortality and reproductive rates as output layers. The Levenberg Marquardt algorithm was used to optimize the network weights. The results showed that the model's prediction accuracy (R ²>0.87, RMSE<0.26) was significantly better than traditional regression methods. Under optimized conditions (C: N (30:1)/Base: Oil (1:2)), the fluctuation ranges of temperature, humidity, and pH were 17-25 ℃, 55% -75%, and 6.0-8.5, respectively. Meanwhile, the BP prediction model established in this study achieved an average accuracy of 82.53% and 81.28% respectively after 200 iterations in the Soil Ratio and C: N scenarios. This study achieved intelligent optimization and quality standardization control of earthworm farming conditions through BP neural network, providing a scientific basis for the standardized production of animal based traditional Chinese medicinal materials.