On the Distinction between Picea koraiensis Nak. and Picea abies(L.) Karsten based on the Discriminant Function (I)

판별식(判別式)에 의한 수목분류법(樹木分類法)에 관(關)하여 (I) -독일(獨逸)가문비와 종비(樅榧)나무와의 판별분석(判別分析)-

This experiment was carried out to distinguish between picea abies (L.) Karsten and Picea koraiensis Nak by the method of discriminant analysis which is used the metrical continuous characteristic on current inorphological plant taxanomy. The results are summarized as follows 1) The discriminant function and discriminant region from the experiment are Z(x)=Z($x_1,\;x_2$)=$0.000379x_1+0.004354x_2-0.311061$ or Z(x)=Z($x_1,\;x_2$=$0.000379(x_1-60.442800)+0.004354(x_2-66.185100)$, $$R_1=(x{\mid}0.000379x_1+0.004354x_2-0.311061{\geq_-}0)$$, $R_2$=($x{\mid}0.000379x_1+0.004354x_2-0.311061$ <0). 2) The probability of misclassification based on the above discriminant region is P($2{\mid}1$)=$P(1{\mid}2)$=0.444 therefore the probability of simultaneous misclassification of P($2{\mid}1$) and $P(1{\mid}2)$ is about 44.4%. 3) the probability of misclassification by the discriminant function resulted from the experiment if recorded as high but it is thought that there is a considerable meaning to perceive the probability of confidence about the discrimination better than its precision.

현행(現行)의 형태학적수목분류법(形態學的樹木分類法)에 계획적특성(計劃的特性)을 이용(利用)한 소위(所謂) 판별분석법(判別分析法)에 의(依)할 독일(獨逸)가문비(Picea abies(L.) Karsten)와 종비나무(樅榧)(Picea koraiensis Nak.)와의 분류시험(分類試驗)을 실시(實施)하고 그 결과(結果)를 다음과 같이 요약(要約)한다. 1) 본(本) 시험(試驗)에서 얻은 판별식(判別式)과 판별영역(判別領域)은, Z(x)=Z($x_1,\;x_2$)=$0.000379x_1+0.004354x_2-0.311061$ 또는 Z(x)=Z($x_1,\;x_2$)=$0.000379(x_1-60.4428)+0.004354(x_2-66.1851)$, $$R_1=\{x{\mid}0.000379x_1+0.004354x_2-0.311061{\geq_-}0\}$$, $R_2$={$x{\mid}0.000379x_1+0.004354x_2-0.311061$ <0)} 또는 $$R_1=\{x{\mid}0.000379(x_1-60.4428)+0.004354(x_2-66.1851){\geq_-}0%\}$$, $R_2$={$x{\mid}0.000379(x_1-60.4428)+0.004354(x_2-66.1851)$ <0}. 2) 위의 판별영역(判別領域)에 의(依)한 오판율(誤判率)(오분류확율(誤分類確率))은, P($2{\mid}1$)=($P1{\mid}2$)=0.444로서, P($2{\mid}1$)와 P($P1{\mid}2$)의 동시오판율(同時誤判率)은 P=44.4%. 3) 본(本) 시험(試驗)에서 얻은 판별식(判別式)에 의(依)한 오판율(誤判率)은 상당(相當)히 높게 나타났지만 그의 정도(精度)보다는 오히려 판별(判別)에 대(對)한 신뢰도(信賴度)를 알 수 있다는데 보다 큰 의의(意義)가 있는 것으로 사료(思料)된다.