In computer science, a logistic model tree (LMT) is a classification model with an associated supervised training algorithm that combines logistic regression (LR) and decision tree learning.[1][2] Logistic model trees are based on the earlier idea of a model tree: a decision tree that has linear regression models at its leaves to provide a piecewise linear regression model (where ordinary decision trees with constants at their leaves would produce a piecewise constant model).[1] In the logistic variant, the LogitBoost algorithm is used to produce an LR model at every node in the tree; the node is then split using the C4.5 criterion. Each LogitBoost invocation is warm-started[vague] from its results in the parent node. Finally, the tree is pruned.[3]

This research introduces a new strategy in cluster ensemble selection by using Independency and Diversity metrics. In recent years, Diversity and Quality, which are two metrics in evaluation procedure, have been used for selecting basic clustering results in the cluster ensemble selection. Although quality can improve the final results in cluster ensemble, it cannot control the procedures of generating basic results, which causes a gap in prediction of the generated basic results' accuracy. Instead of quality, this paper introduces Independency as a supplementary method to be used in conjunction with Diversity. Therefore, this paper uses a heuristic metric, which is based on the procedure of converting code to graph in Software Testing, in order to calculate the Independency of two basic clustering algorithms. Moreover, a new modeling language, which we called as "Clustering Algorithms Independency Language" (CAIL), is introduced in order to generate graphs which depict Independency of algorithms. Also, Uniformity, which is a new similarity metric, has been introduced for evaluating the diversity of basic results. As a credential, our experimental results on varied different standard data sets show that the proposed framework improves the accuracy of final results dramatically in comparison with other cluster ensemble methods.

This paper presents a semi-parametric algorithm for online learning of a robot inverse dynamics model. It combines the strength of the parametric and non-parametric modeling. The former exploits the rigid body dynamics equa- tion, while the latter exploits a suitable kernel function. We provide an extensive comparison with other methods from the literature using real data from the iCub humanoid robot. In doing so we also compare two different techniques, namely cross validation and marginal likelihood optimization, for estimating the hyperparameters of the kernel function.

Our team won the second prize of the Safe Aging with SPHERE Challenge organized by SPHERE, in conjunction with ECML-PKDD and Driven Data. The goal of the competition was to recognize activities performed by humans, using sensor data. This paper presents our solution. It is based on a rich pre-processing and state of the art machine learning methods. From the raw train data, we generate a synthetic train set with the same statistical characteristics as the test set. We then perform feature engineering. The machine learning modeling part is based on stacking weak learners through a grid searched XGBoost algorithm. Finally, we use post-processing to smooth our predictions over time.

Machine learning is kind of magic right? But is it the kind of magic that can make us rich? And I don't mean lucrative consulting gig rich, I mean digging valuable metals out of the ground rich. Also I'd been meaning to try out some transfer learning and looking around for a good topic to try it on. Transfer learning is where you take a pre-trained convolution (or other) network and use it for your task.

Have people not been doing this up till now? Seems fairly obvious to me. A similar technique I've used for fun is applying k-NN to each layer to see how well the classes are separated at this layer based on the Euclidean distance (or whatever other metric one wishes to employ). Note that technically k-NN doesn't need any training at all!

Professors Hastie and Tibshirani published "The Elements of Statistical learning: Data mining, inference and prediction", with Jerome Friedman (springer, 2001, second edition 2009). This book has received a terrific reception, with over 45,000 copies sold. Both presenters are actively involved in research in statistical learning methods, and are well-known not only in the statistics community but in the machine-learning, neural network and bioinformatics fields as well. Their newer book "An Introduction to Statistical Learning, with Applications in R" (with Gareth James and Daniela Witten, 2013) is also a best-seller, and has remained consistently in the top 10 in the Amazon categories "Mathematics and Statistics" and "Artificial Intelligence", with a five-star rating based on 84 customer reviews. Over the years they have become leaders in the statistical analysis of DNA microarrays, working with leading-edge biologists such as Patrick Brown of Stanford University, and David Botstein of Princeton.

We consider the online version of the isotonic regression problem. Given a set of linearly ordered points (e.g., on the real line), the learner must predict labels sequentially at adversarially chosen positions and is evaluated by her total squared loss compared against the best isotonic (non-decreasing) function in hindsight. We survey several standard online learning algorithms and show that none of them achieve the optimal regret exponent; in fact, most of them (including Online Gradient Descent, Follow the Leader and Exponential Weights) incur linear regret. We then prove that the Exponential Weights algorithm played over a covering net of isotonic functions has a regret bounded by $O\big(T^{1/3} \log^{2/3}(T)\big)$ and present a matching $\Omega(T^{1/3})$ lower bound on regret. We provide a computationally efficient version of this algorithm. We also analyze the noise-free case, in which the revealed labels are isotonic, and show that the bound can be improved to $O(\log T)$ or even to $O(1)$ (when the labels are revealed in isotonic order). Finally, we extend the analysis beyond squared loss and give bounds for entropic loss and absolute loss.

Non-parametric estimation of a multivariate density estimation is tackled via a method which combines traditional local smoothing with a form of global smoothing but without imposing a rigid structure. Simulation work delivers encouraging indications on the effectiveness of the method. An application to density-based clustering illustrates a possible usage.

Detecting emergence of a low-rank signal from high-dimensional data is an important problem arising from many applications such as camera surveillance and swarm monitoring using sensors. We consider a procedure based on the largest eigenvalue of the sample covariance matrix over a sliding window to detect the change. To achieve dimensionality reduction, we present a sketching-based approach for rank change detection using the low-dimensional linear sketches of the original high-dimensional observations. The premise is that when the sketching matrix is a random Gaussian matrix, and the dimension of the sketching vector is sufficiently large, the rank of sample covariance matrix for these sketches equals the rank of the original sample covariance matrix with high probability. Hence, we may be able to detect the low-rank change using sample covariance matrices of the sketches without having to recover the original covariance matrix. We character the performance of the largest eigenvalue statistic in terms of the false-alarm-rate and the expected detection delay, and present an efficient online implementation via subspace tracking.