Improvement of the Data Analysis Algorithm by Applying the Decision Tree Method.
Jung, Won Chan · 1992
Data analysis (reconstructability analysis) is an area used on a data set which has several variables and a function value to find the most important factor that causes the function values to fall within a desired range. Normal data analysis algorithm (1) finds the most important factor in O (2$\sp{n})$ time. This dissertation introduces a newly developed system of algorithms called the Data Analyzing Tree (DAT) which is designed to either reduce the time complexity or produce more accurate results. DAT-1 uses O $(n\sp2)$ time to produce the same results as the normal data analysis method, and DAT-2 produces the result with a higher fall-into-the-range rate while using the same time complexity as the normal data analysis. Therefore, DAT-1 is suitable to get quick results, and DAT-2 or a higher numbered DAT is suitable to get more accurate results. DATs give more choices of the algorithm, so the users can choose the appropriate algorithm depending on the circumstances.