A soil heavy metal content prediction Based Stochastic-driven Sparrow Search Algorithm Optimized Least Squares Support Vector Machine model

To reduce prediction errors for soil heavy metal content prediction, Stochastic Sparrow Search Algorithm Optimized Least Squares Support Vector Machine (Rd-SSA-LSSVM) optimization model is proposed. Stochastic-driven Sparrow Search Algorithm (Rd-SSA) is used to find the parameters of Least Squares Support Vector Machine (LSSVM). A stochastic strategy was introduced in Rd-SSA to de-influence sparrow finder, hazard and best sparrow location updates to balance local exploitation and global search capabilities. Combined with the prediction of heavy metal Cr in agricultural soils in East-West Lake District, Wuhan, Hubei Province (2019), it has been shown that better prediction performance and generalisation performance is found in Rd-SSA-LSSVM, rather than the other comparison models.