Importance

Loading Importance Indicators...

Abstract

Computing technologies for benchmarking, which may be based on a k-Nearest Neighbor (kNN) algorithm or another suitable machine learning algorithm, solve a cold start problem for a recommender engine employing a collaborative filtering algorithm when training data of user preferences or actual data of user actions is not available. For example, the cold start problem's unavailability of labeled data to train and develop a supervised model may be addressed by breaking the cold start problem down into two parts. The first part includes a KNN (or another suitable algorithm) model to cluster profiles based on a set of variables and this implementation of the KNN model is unsupervised, since there is no labeled data available to train the KNN model. The second part includes a prioritization algorithm that leverages certain outputs from the KNN model and neighbors to prioritize benchmarks. As such, using this two part approach, the cold start problem is solved by leveraging an unsupervised model to address lack of the labeled data.

Loading the Abstract Image...

First Claim