鳜塘水体生态因子的计算机模式识别法应用

Application of the ecosystem factors in pond waters for Siniperca chuatsi by computer pattern recognition

应用计算机模式识别技术研究鳜塘水体中溶氧量与九项生态因子间的相互关系。研究表明 ,不同塘或不同时期其生态因子差异较大。在高维空间中 ,高、低溶氧量两类样品点聚集在不同的区域 ,周界清晰、分类十分成功 ,表明池塘水体溶氧量与水中多项生态因子间存在显著规律性。用模式识别优化技术求得维持稳定的高溶氧状态的 9项生态指标值 ,优化生态因子的溶氧预报值大于 8.4mg/L。水体溶氧量与其它九项因子成多元的二次函数关系 ,其模型复相关系数R =0 .97。对测量点溶氧量的回代准确率为 93.9%。

The computer pattern recognition(CPR) technology is applied to the study on the relationships among dissolved oxygen (DO) concentration and nine ecosystem factors in the Siniperca chuatsi ponds. The factors differ greatly in distinct ponds or periods . In high-dimensional space, it is clear and successful that two kinds of samples, high and low DO, gather themselves in different regions. It indicates markedly regular relationships between DO and ecosystem factors. The nine superiority ecosystem targets for maintain high DO are sought by CPR favorable technology. In water, DO and nine ecosystem factors follow polybasic quadratic function relation, R=0.97. Using the mathematical model prediction DO, precision is 93.9%, predicting the DO is over 8.4 mg/L for favorable ecosystem factor.