First page Back Continue Last page Overview Text


Future works will include a mathematical approach to further mathematically explain the behaviours of coefficients and a machine learning approach to data fusion based on combining the best coefficients from each retrieved-actives size range. This also includes a study on whether when given a query size, the size distribution of coefficients behave the same regardless the active type. So far, we compared the size distributions of classese, antihypertensive and hypolipidemic. When given the same query size, their size distributions of retrieved active are very similar although not exactly same.