中文摘要：

对50例非癌症患者和100例癌症患者分组，运用邻近点插值的数学模型诊断。对诊断对象的腺苷三磷酸酶（ATP酶）和琥珀酸脱氢酶（SOH酶）活性两项指标利用邻近点插值，其诊断正确率可以高达98．2％，增加训练样本集的覆盖面，诊断正确率将趋于100％。邻近点插值是癌症诊断的高效数学模型。

英文摘要：

Fifty healthy people and 100 cancer patients were diagnosed with the mathematical model of nearest neighbor interpolation. The correct diagnosis rate of cancer was up to 98.2% by using nearest neighbor interpolation to process the values of adenosine triphosphatase （ATPase） and succinate dehydrogenase （SDHase） of all the subjects. It could get very close to 100%, if the coverage of training sample set was enlarged. Nearest neighbor interpolation is an efficacions mathematical model in cancer diagnosis.