Multi-instance clustering with applications to multi-instance prediction

Abstract

Abstract. In the setting of multi-instance learning, each object is represented by a bag composed of multiple instances instead of by a single instance in traditional learning setting. Previous works in this area only concern multi-instance prediction problems where each bag is associated with a binary (classification) or real-valued (regression) label. However, unsupervised multi-instance learning where bags are without labels have not been studied. In this paper, the problem of unsupervised multiinstance learning is addressed where a multi-instance clustering algorithm named Bamic is proposed.Briefly, by regarding bags as atomic data items and using some form of distance metric to measure distances between bags, Bamic adapts the popular k -Medoids algorithm to partition the unlabeled training bags into k disjoint groups of bags. Furthermore, based on the clustering results, a novel multi-instance prediction algorithm named Bartmip is developed. Firstly, each bag is re-represented by a k -dimensional feature vector, where the value of the i -th feature is set to be the distance between the bag and the medoid of the i -th group. After that, bags are transformed into feature vectors so that common supervised learners are used to learn from the transformed feature vectors each associated with the original bag's label. Extensive experiments show that Bamic could effectively discover the underlying structure of the data set and Bartmip works quite well on various kinds of multi-instance prediction problems.

Mining Ticketing Logs for Usage Characterization with Nonnegative Matrix Factorization

“…As we extract 100 atoms per frequency band and there are around 300 stations in our problem, these matrices have small dimensions. We use a multi-scale clustering algorithm, similar to what [25], to cluster together stations into coherent groups of behaviors. Since the matrices are small (300 × 100), the clustering is fast (a couple of seconds) on a typical computer.…”

Reducing Dimensionality in Multiple Instance Learning with a Filter Method

“…Therefore, a new concept has to be introduced concerning the similarity function and the calculation of the differences between attribute A in patterns R and H, dif f (A, R, H). The literature proposes different distance-based approaches to solve MI problems [8,9,10]. The most extensively used metric is Hausdorff distance [11], which measures the distance between two sets.…”

“…This extension of ReliefF for MIL uses the Average Hausdorff distance [10] proposed by Zhang and Zhou to measure the distance between two bags. It is defined as follows:…”

An improved MIMLRBF natural scene image classification based on spectral clustering

“…Neural network is a kind of practical techniques in machine learning, and it plays a role in the problem of MIML. Zhang M-L [7] proposed MIMLRBF(Multi-Instance Multi-Label Radial Basis Function) based on neural network, the algorithm made full use of the relationships between instances and labels and get better effects. The key of the MIMLRBF algorithm is to obtain clustering center.…”

“…The key of the MIMLRBF algorithm is to obtain clustering center. Another key of the MIMLRBF algorithm is the measure way of the distances between two packages, and the performance of algorithm is improved when improving the distances between two packages in the literature [6] and literature [7].…”