Abstract: The least square support vector machine(LS-SVM)was used to establish an analytical prediction model of ammonia nitrogen concentration as a key water quality index in aquaculture in order to improve the measurement accuracy of ammonia nitrogen concentration,and the analytical prediction model was optimized to reduce the influence of pH value,temperature,static time and otherfactors during measurement process by an orthogonal simulation test.The results showed that the precise measurement of ammonia nitrogen concentration was observed under the conditions of pH value of 10.5,temperature of 35℃,static time of 20 min,and light source wave length of 380 nm using an on-line ammonia nitrogen detection system.
居锦武. 基于LS-SVM的养殖水体氨氮含量分析模型的优化[J]. 大连海洋大学学报, 2016, 31(4): 444-448.
JU Jin-wu. Optimization of an ammonia nitrogen content analysis model in aquaculture water based on LS-SVM. Journal of Dalian Ocean University, 2016, 31(4): 444-448.