@inproceedings{oai:jaxa.repo.nii.ac.jp:00020372, author = {杉田, 精司 and 廣井, 孝弘 and 永田, 賢二 and 岡田, 真人 and Tsuboi, Nobuko and Sugita, Seiji and Hiroi, Takahiro and Nagata, Kenji and Okada, Masato}, book = {月・惑星シンポジウム, Proceedings of the ISAS lunar and planetary symposium}, month = {Aug}, note = {第42回 月・惑星シンポジウム(2009年8月5-7日, 宇宙航空研究開発機構宇宙科学研究本部), The 42nd ISAS Lunar and Planetary Symposium (August 5-7, 2009. Institute of Space and Astronautical Science, Japan Aerospace Exploration Agency (JAXA)), Sagamihara, Kanagawa Japan, Modified Gaussian model (MGM) is a widely used analytical method for reflectance spectroscopy of planetary surfaces in the visible and near-infrared range. One of the current challenges of the MGM is its automation to handle a large quantity of spectral data. In this study, we propose a new MGM scheme by adopting an approach called the cross-validation method in order to automatically determine the optimum number of Gaussians. As the first step of this study, we searched for the optimum number of Gaussians in reflectance spectra of olivine using our proposed method. MGM analyses with cross-validation tests indicate that olivine spectra are separated into four Gaussians to yield optimum results. The obtained optimum number (four) of Gaussians is the same as that predicted from well-controlled single crystal measurements, suggesting that cross-validation method can find the optimum number of absorption band accurately. Furthermore, MGM analyses on reflectance spectra of olivine powders with a variety of Fe/Mg ratio using four Gaussians yield a very different relationship between olivine composition and Gaussian parameters from those by previous analyses, underlining the importance of the choice of the number of Gaussians in MGM analyses., 形態: カラー図版あり, 形態: CD-ROM1枚, Physical characteristics: Original contains color illustrations, Note: One CD-ROM, 資料番号: AA0064739028}, pages = {118--122}, publisher = {宇宙航空研究開発機構宇宙科学研究本部, Institute of Space and Astronautical Science, Japan Aerospace Exploration Agency (JAXA)}, title = {Development of a new modified Gaussian model (MGM) scheme using the cross validation method}, volume = {42}, year = {2009} }