Multiple kernel learning

About: Multiple kernel learning is a research topic. Over the lifetime, 1630 publications have been published within this topic receiving 56082 citations.

Efficient Audio Recognition Algorithm based on Simple Multiple Kernel Learning

Abstract: On account of limitations and shortcomings of traditional audio recognition model, audio recognition with low SNR is deeply searched in this paper. Considering the functions and features of audio recognition, the general steps of audio recognition are analyzed and the application of Simple Multiple Kernel Learning (SMKL) in audio recognition with low SNR is presented to improve the recognition rate and accuracy of audio. The experimental results show that SMKL has a higher accuracy in identifying audio under the circumstance of low SNR than that recognition rate of each time of SMKL algorithm is higher than that of SVM algorithm. SMKL can be well applied to circumstances of large-scale sample data, complex dimension and massive heterogeneous information. Accuracy of audio recognition of kernel parameters optimization with grid-search method is higher than that with the method of fixed kernel parameters, the accuracy can be up to 85.52%. What’s more, effectiveness of grid-search method when determining kernel parameters can be seen from classification results

Averaging of kernel functions

Abstract: In kernel-based machines, the integration of several kernels to build more flexible learning methods is a promising avenue for research. In particular, in Multiple Kernel Learning a compound kernel is build by learning a kernel that is the weighted mean of several sources. We show in this paper that the only feasible average for kernel learning is precisely the arithmetic average. We also show that three familiar means (the geometric, inverse root mean square and harmonic means) for positive real values actually generate valid kernels.

Semi-supervised cross-view scene model adaptation for remote sensing image classification

Abstract: In this paper, we address the problem of semi-supervised visual domain adaptation for transferring scene category models from ground view images to overhead view very high-resolution (VHR) remote sensing images. We introduce a multiple kernel learning domain adaptation algorithm to fuse the information from multiple features and cope with the considerable variation in feature distributions between images from two domains. For each image, we first extract eight state-of-art local features and use the pretrained scene attribute model from ground-level SUN attribute database to predict attribute labels. For each scene class we learn an adapted target classifier based on multiple feature kernels by minimizing both the structural risk functional and the mismatch between data distributions of two domains. Experimental results demonstrate that it is possible to use a scene category model learned on a set of ground view scenes for semi-supervised classification of VHR remote sensing images.

A multimodality approach to predicting the popularity of sneakers

Abstract: We present a computational approach for predicting the popularity score of sneakers through the analysis of growing amount of online data Sneakers are described in several aspects based on which a popularity prediction model is constructed In particular, we utilize the multiple kernel learning technique with customized kernels to analyze multimodal data extracted from an online sneaker magazine The construction of a prediction model from multiple facets is not trivial — the effectiveness of each feature depends on the way we compute and combine it with the others We examine a few design choices and study how multimodal data should be utilized to achieve practical prediction

Vehicle Travel Time Predication based on Multiple Kernel Regression

Abstract: With the rapid development of transportation and logistics economy, the vehicle travel time prediction and planning become an important topic in logistics. Travel time prediction, which is indispensible for traffic guidance, has become a key issue for researchers in this field. At present, the prediction of travel time is mainly short term prediction, and the predication methods include artificial neural network, Kaman filter and support vector regression (SVR) method etc. However, these algorithms still have some shortcomings, such as highcomputationcomplexity, slow convergence rate etc. This paper exploits the learning ability of multiple kernel learning regression (MKLR) in nonlinear prediction processing characteristics, logistics planning based on MKLR for vehicle travel time prediction. The method for Vehicle travel time prediction includes the following steps: (1) preprocessing historical data; (2) selecting appropriate kernel function, training the historical data and performing analysis ;(3) predicting the vehicle travel time based on the trained model. The experimental results show that, through the analysis of using different methods for prediction, the vehicle travel time prediction method proposed in this paper, archives higher accuracy than other methods. It also illustrates the feasibility and effectiveness of the proposed prediction method.