Real-world social networks are dynamic in nature. Companies continue to collaborate, align strategically, acquire, and merge over time, and receive positive/negative impact from other companies. Consequently, their performance changes with time. If one can understand what types of network changes affect a company's value, he/she can predict the future value of the company, grasp industry innovations, and make business more successful. However, it often requires continuous records of relational changes, which are often difficult to track for companies, and the models of mining longitudinal network are quite complicated. In this study, we developed algorithms and a system to infer large-scale evolutionary company networks from public news during 1981--2009. Then, based on how networks change over time, as well as the financial information of the companies, we predicted company profit growth. This is the first study of longitudinal network-mining-based company performance analysis in the literature.

The rapid construction of supervised text classification models is becoming a pervasive need across many modern applications. To reduce human-labeling bottlenecks, many new statistical paradigms (e.g., active, semi-supervised, transfer and multi-task learning) have been vigorously pursued in recent literature with varying degrees of empirical success. Concurrently, the emergence of Web 2.0 platforms in the last decade has enabled a world-wide, collaborative human effort to construct a massive ontology of concepts with very rich, detailed and accurate descriptions. In this paper we propose a new framework to extract supervisory information from such ontologies and complement it with a shift in human effort from direct labeling of examples in the domain of interest to the much more efficient identification of concept-class associations. Through empirical studies on text categorization problems using the Wikipedia ontology, we show that this shift allows very high-quality models to be immediately induced at virtually no cost.

Fraud detection is a key activity with serious socio-economical impact. Inspection activities associated with this task are usually constrained by limited available resources. Data analysis methods can provide help in the task of deciding where to allocate these limited resources in order to optimise the outcome of the inspection activities. This paper presents a multi-strategy learning method to address the question of which cases to inspect first. The proposed methodology is based on the utility theory and provides a ranking ordered by decreasing expected outcome of inspecting the candidate cases. This outcome is a function not only of the probability of the case being fraudulent but also of the inspection costs and expected payoff if the case is confirmed as a fraud. The proposed methodology is general and can be useful on fraud detection activities with limited inspection resources. We experimentally evaluate our proposal on both an artificial domain and on a real world task.

Data streams containing objects that are (or can be) associated with more than one label at the same time are ubiquitous. In spite of its important applications, classification of streaming multi-label data is largely unexplored. Existing approaches try to tackle the problem by transferring traditional single-label stream classification practices to the multi-label domain. Nevertheless, they fail to consider some of the unique properties of the problem such as within and between class imbalance and multiple concept drift. To deal with these challenges, this paper proposes a novel multi-label stream classification approach that employs two windows for each label, one for positive and one for negative examples. Instance-sharing is exploited for space efficiency, while a time-efficient instantiation based on the k-Nearest Neighbor algorithm is also proposed. Finally, a batch-incremental thresholding technique is proposed to further deal with the class imbalance problem. Results of an empirical comparison against two other methods on three real world datasets are in favor of the proposed approach.

Understanding extreme events, such as hurricanes or forest fires, is of paramount importance because of their adverse impacts on human beings. Such events often propagate in space and time. Predicting-even a few days in advance-what locations will get affected by the event tracks could benefit our society in many ways. Arguably, simulations from “first principles,” where underlying physics-based models are described by a system of equations, provide least reliable predictions for variables characterizing the dynamics of these extreme events. Data-driven model building has been recently emerging as a complementary approach that could learn the relationships between historically observed or simulated multiple, spatio-temporal ancillary variables and the dynamic behavior of extreme events of interest. While promising, the methodology for predictive learning from such complex data is still in its infancy. In this paper, we propose a dynamic networks-based methodology for in-advance prediction of the dynamic tracks of emerging extreme events. By associating a network model of the system with the known tracks, our method is capable of learning the recurrent network motifs that could be used as discriminatory signatures for the event's behavioral class. When applied to classifying the behavior of the hurricane tracks at their early formation stages in Western Africa region, our method is able to predict whether hurricane tracks will hit the land of the North Atlantic region at least 10-15 days lead lag time in advance with more than 90% accuracy using 10-fold cross-validation. To the best of our knowledge, no comparable methodology exists for solving this problem using data-driven models.

The mixture of multivariate Bernoulli distributions (MMB) is a statistical model for high-dimensional binary data in widespread use. Recently, the MMB has been used to model the sequence of packet receptions and losses of wireless links in sensor networks. Given an MMB trained on long data traces recorded from links of a deployed network, one can then use samples from the MMB to test different routing algorithms for as long as desired. However, learning an accurate model for a new link requires collecting from it long traces over periods of hours, a costly process in practice (e.g. limited battery life). We propose an algorithm that can adapt a preexisting MMB trained with extensive data to a new link from which very limited data is available. Our approach constrains the new MMB's parameters through a nonlinear transformation of the existing MMB's parameters. The transformation has a small number of parameters that are estimated using a generalized EM algorithm with an inner loop of BFGS iterations. We demonstrate the efficacy of the approach using the MNIST dataset of handwritten digits, and wireless link data from a sensor network. We show we can learn accurate models from data traces of about 1 minute, about 10 times shorter than needed if training an MMB from scratch.

In order to deal with ambiguity in natural language, it is common to organise words, according to their senses, in synsets, which are groups of synonymous words that can be seen as concepts. The manual creation of a broad-coverage synset base is a time-consuming task, so we take advantage of dictionary definitions for extracting synonymy pairs and clustering for identifying synsets. Since word senses are not discrete, we create fuzzy synsets, where each word has a membership degree. We report on the results of the creation of a fuzzy synset base for Portuguese, from three electronic dictionaries. The resulting resource is larger than existing hancrafted Portuguese thesauri.

The problem of feature selection has aroused considerable research interests in the past few years. Traditional learning based feature selection methods separate embedding learning and feature ranking. In this paper, we introduce a novel unsupervised feature selection approach via Joint Embedding Learning and Sparse Regression (JELSR). Instead of simply employing the graph laplacian for embedding learning and then regression, we use the weight via locally linear approximation to construct graph and unify embedding learning and sparse regression to perform feature selection. By adding the ell {2,1} -norm regularization, we can learn a sparse matrix for feature ranking. We also provide an effective method to solve the proposed problem. Compared with traditional unsupervised feature selection methods, our approach could integrate the merits of embedding learning and sparse regression simultaneously. Plenty of experimental results are provided to show the validity.

Employing different We investigate new methods for creating and applying coreference models to create ensembles bears resemblance ensembles for coreference resolution. While to Pang and Fan's [2009] approach, where an ensemble of existing ensembles for coreference resolution are pairwise models is applied to Chinese coreference resolution, typically created using different learning algorithms, but contrasts with the vast majority of existing approaches, clustering algorithms or training sets, we where an ensemble of coreference systems is typically created harness recent advances in coreference modeling by employing different learning algorithms [Munson et and propose to create our ensemble from a variety al., 2005] or clustering algorithms [Ng, 2005], or perturbing of supervised coreference models. However, the training set using meta-learning techniques such as the presence of pairwise and non-pairwise coreference bagging and boosting [Ng and Cardie, 2003; Kouchnir, 2004; models in our ensemble presents a challenge Vemulapalli et al., 2009].

There are two key issues for information diffusion in blogosphere: (1) blog posts are usually short, noisy and contain multiple themes, (2) information diffusion through blogosphere is primarily driven by the “word-of-mouth” effect, thus making topics evolve very fast. This paper presents a novel topic tracking approach to deal with these issues by modeling a topic as a semantic graph in which the semantic relatedness between terms are learned from Wikipedia. For a given topic/post, the named entities, Wikipedia concepts, and the semantic relatedness are extracted to generate the graph model. Noises are filtered out through a graph clustering algorithm. To handle topic evolution, the topic model is enriched by using Wikipedia as background knowledge. Furthermore, graph edit distance is used to measure the similarity between a topic and its posts. The proposed method is tested using real-world blog data. Experimental results show the advantage of the proposed method on tracking topics in short, noisy text.