Matthieu Exbrayat

CoFKM: A Centralized Method for Multiple-View Clustering

This paper deals with clustering for multi-view data, i.e. objects described by several sets of variables or proximity matrices. Many important domains or applications such as Information Retrieval, biology, chemistry and marketing are concerned by this problematic. The aim of this data mining research field is to search for clustering patterns that perform a consensus between the patterns from different views. This requires to merge information from each view by performing a fusion process that identifies the agreement between the views and solves the conflicts. Various fusion strategies can be applied, occurring either before, after or during the clustering process. We draw our inspiration from the existing algorithms based on a centralized strategy. We propose a fuzzy clustering approach that generalizes the three fusion strategies and outperforms the main existing multi-view clustering algorithm both on synthetic and real datasets.

PIPELINING A SKEW-INSENSITIVE PARALLEL JOIN ALGORITHM

Most standard parallel join algorithms try to overcome data skews with a relatively static approach. The way they distribute data (and then computation) over nodes depends on a data re-distribution algorithm (hashing or range partitioning) that is determined before the actual join begins. On the contrary, our approach consists in pre-scanning data in order to choose an efficient join method for each given value of the join attribute. This approach has already proved to be efficient both theoretically and practically in our previous papers. In this paper we introduce a new pipelined version of our frequency adaptive join algorithm. The use of pipelining offers flexible strategies for resource allocation while avoiding unnecessary disk input/output of intermediate join results when computing multi-join queries. We present a detailed version of the algorithm and a cost analysis based on the BSP (Bulk Synchronous Parallel) model, showing that our pipelined algorithm achieves noticeable improvements compared to the sequential parallel version for multi-join queries while guaranteeing perfect balancing properties.

Generative structure learning for Markov logic networks based on graph of predicates

In this paper we present a new algorithm for generatively learning the structure of Markov Logic Networks. This algorithm relies on a graph of predicates, which summarizes the links existing between predicates and on relational information between ground atoms in the training database. Candidate clauses are produced by means of a heuristical variabilization technique. According to our first experiments, this approach appears to be promising.

A PC-NOW Based Parallel Extension for a Sequential DBMS

In this paper we study the use of networks of PCs to handle the parallel execution of relational database queries. This approach is based on a parallel extension, called parallel relational query evaluator, working in a coupled mode with a sequential DBMS. We present a detailed arc hitecture of the parallel query evaluator and introduce Enkidu, the effiient Java-based prototype that has been build according to our concepts. We expose a set of measurements, conducted over Enkidu, and highlighting its performances.We finally discuss the interest and viability of the concept of parallel extension in the context of relational databases and in the wider context of high performance computing.

Integrating Pairwise Constraints into Clustering Algorithms: Optimization-Based Approaches

In this paper we introduce new models for semi-supervised clustering problem, in particular we address this problem from the representation space point of view. Given a dataset enhanced with constraints (typically must-link and cannot-link constraints) and any clustering algorithm, the proposed approach aims at learning a projection space for the dataset that satisfies not only the constraints but also the required objective of the clustering algorithm on unenhanced data. We propose a boosting framework to weight the constraints and infers successive projection spaces in such a way that algorithm performance is improved. We experiment this approach on standard UCI datasets and show the effectiveness of our algorithm.

Discriminative Markov logic network structure learning based on propositionalization and X 2 -test

In this paper we present a bottom-up discriminative algorithm to learn automatically Markov Logic Network structures. Our approach relies on a new propositionalization method that transforms a learning dataset into an approximative representation in the form of boolean tables, from which to construct a set of candidate clauses according to a χ2-test. To compute and choose clauses, we successively use two different optimization criteria, namely pseudo-log-likelihood (PLL) and conditional log-likelihood (CLL), in order to combine the efficiency of PLL optimization algorithms together with the accuracy of CLL ones. First experiments show that our approach outperforms the existing discriminative MLN structure learning algorithms.

Interactive and progressive constraint definition for dimensionality reduction and visualization

Projecting and visualizing objects in a two- or tree-dimension space is a standard data analysis task. In addition to this visualization it might be of interest to allow the user to add knowledge in the form of (di)similarity constraints among objects, when those appear either too close or too far in the observation space. In this paper we propose three kinds of constraints and present a resolution method that derives from PCA. Experiments have been performed with both synthetic and usual datasets. They show that a relevant representation can be achieved with a limited set of constraints.

Time series symbolization and search for frequent patterns

In this paper, we focus on two aspects of time series mining: first on the transformation of numerical data to symbolic data; then on the search for frequent patterns in the resulting symbolic time series. We are thus interested in some patterns which have a high frequency in our database of time series and might help to generate candidates for various tasks in the area of time series mining. During the symbolization phase, we transform the numerical time series into a symbolic time series by i) splitting this latter into consecutive subsequences, ii) using a clustering algorithm to cluster these subsequences, each subsequence being then replaced by the name of its cluster to produce the symbolic time series. In the second phase, we use a sliding window to create a collection of transactions from the symbolic time series, then we use some algorithm for mining sequential pattern to find out some interesting motifs in the original time series. An example experiment based on environmental data is presented.

Heuristic Method for Discriminative Structure Learning of Markov Logic Networks

In this paper, we present a heuristic-based algorithm to learn discriminative MLN structures automatically, directly from a training dataset. The algorithm heuristically transforms the relational dataset into boolean tables from which it builds candidate clauses for learning the final MLN. Comparisons to the state-of-the-art structure learning algorithms for MLNs in the three real-world domains show that the proposed algorithm outperforms them in terms of the conditional log likelihood (CLL), and the area under the precision-recall curve (AUC).

Automatic classification of ceramic sherds with relief motifs

Abstract. A large corpus of ceramic sherds dating from the High Middle Ages has been extracted in Saran (France). The sherds have an engraved frieze made by the potter with a carved wooden wheel. These relief patterns can be used to date the sherds in order to study the diffusion of ceramic production. The aim of the ARCADIA project was to develop an automatic classification of this archaeological heritage. The sherds were scanned using a three-dimensional (3-D) laser scanner. After projecting the 3-D point cloud onto a depth map, the local variance highlighted the shallow relief patterns. The saliency region focused on the motif was extracted by a density-based spatial clustering of FAST points. An adaptive thresholding was then applied to the depth to obtain a binary pattern close to manual sampling. The five most representative types of motif were classified by training an SVM model with a pyramid histogram of visual words descriptor. Compared with other state-of-the-art methods, the proposed approach succeeded in classifying up to 84% of the binary patterns on a dataset of 377 scanned sherds. The automatic method is extremely time-saving compared to manual stamping.