Classification Model of Ancient Glass Artifacts Using Fisher Discriminant Analysis

Ancient glass artifacts are highly susceptible to weathering in burial environments, which can cause changes their chemical composition. To address the problem of identifying and analysing composition ancient artifacts, this paper proposes a classification identification model based on Fisher discriminant method. Firstly, considering that proportion components change during process, affect correct its category, predicts data before normal distribution law performs closed operations ensure sum component is 100%. Secondly, through variance analysis kernel density estimation, factors affecting selected, linear established. Finally, cross-validation performed by applying both training testing data, results two groups found be consistent with actual situation, discrimination accuracy rate