@techreport{oai:jaxa.repo.nii.ac.jp:00044545,
 author = {木村, 武雄 and KIMURA, Takeo},
 month = {Aug},
 note = {The purpose of the theory in this paper is to show how the observation systems should be defined as“better”or“best”quantitatively. On the basis of this theory, we can find the best observation system by means of only a numerical calculation. It should be noted that the word“better”or“best”is not meant economically, socially or so on but has the meaning of certainty or accuracy. In the case in which the unknown state to be obtained by observation is represented by only one variable, the certainty or accuracy of the state is obvious, but in cases in which the unknown state is represented by many variables, the certainty or accuracy is ambiguous in the recent theories. The theory in this report is effective also in the latter case, and we shall examine in a few examples in this report whethere the theory is useful and admissible. The certainty or accuracy in the case of many variables is usually represented by the evaluation functions of quadratic form. However it would be incorrect, as is mathematically verified in this report if the evaluation function should be represented by the so-called generalized variance, which is not of quadratic but of determinant form., 資料番号: NALTR0301000, レポート番号: NAL TR-301},
 title = {観測系の評価に関するひとつの数学的理論},
 year = {1972}
}