This paper discusses the effect of hubness in zero-shot learning, when ridge regression is used to find a mapping between the example space to the label space. Contrary to the existing approach, which attempts to find a mapping from the example space to the label space, we show that mapping labels into the example space is desirable to suppress the emergence of hubs in the subsequent nearest ne...

One of the current challenges concerning improving recommender systems consists of finding ways of increasing serendipity and diversity, without compromising the precision and recall of the system. One possible way to approach this problem is to complement a standard recommender by another recommender “orthogonal” to the standard one, i.e. one that recommends different items than the standard. ...

Finding Nearest Neighbors efficiently is crucial to the design of any nearest neighbor classifier. This paper shows how Layered Range Trees could be used for efficient nearest neighbor classification. The presented algorithm is simple and finds the nearest neighbor in a logarithmic order. It performs d log n + k distance measures to find the nearest neighbor, where k is a constant that is much ...

Phishing causes many problems in business industry. The electronic commerce and banking such as mobile involves a number of online transaction. In transactions, we have to discriminate features related legitimate phishing websites order ensure security the this study, collected data form phish tank public repository proposed K-Nearest Neighbors (KNN) based model for attack detection. detects th...

We derive the exact expressions for the effective masses of the conduction and all three hole bands at G in the nearest-neighbor sps* model, both with and without the spin-orbit interaction. From these expressions we find that the nearest-neighbor sps* model can usually fit the electron mass better than the lightand heavyhole masses and will not be equally successful across a broad spectrum of ...

This paper analyzes k nearest neighbor classification with training data anonymized using anatomy. Anatomy preserves all data values, but introduces uncertainty in the mapping between identifying and sensitive values. We first study the theoretical effect of the anatomized training data on the k nearest neighbor error rate bounds, nearest neighbor convergence rate, and Bayesian error. We then v...