Improving the item to item algorithm of collaborative filtration method for the development of recommendation systems based on the cosine measure by relevant assessment
Vladimir Kucheruk, Mikhail Hlushko · ScienceRise · 2018
The analysis of comparative results of reference systems on the basis of the Tanimoto correlation coefficient in comparison with the "item to item" algorithm of collaborative filtration with the help of relevant assessment is presented. Data for surveys in the form of users with unique IDs are formed. Algorithm of collaborative filtration is based on a cosine measure, which represents the similarity of objects as a cosine between the vectors of purchases in the matrix of users and objects