Catálogo de publicaciones - libros

This paper proposes a new method for generating multi-dimensional sequential patterns. While the current sequential pattern methods are generating patterns within a single attribute, the proposed method is able to detect them among different attributes. We employ an information theoretic method for generating multi-dimensional sequential patterns with the use of Hellinger entropy measure. A number of theorems are proposed to reduce the computational complexity of the sequential pattern systems. The proposed method is tested on some synthesized transaction databases.

Most real-world datasets are, to a certain degree, skewed. When considered that they are also large, they become the pinnacle challenge in data analysis. More importantly, we cannot ignore such datasets as they arise frequently in a wide variety of applications. Regardless of the analytic, it is often that the effectiveness of analysis can be improved if the characteristic of the dataset is known in advance. In this paper, we propose a novel technique to preprocess such datasets to obtain this insight. Our work is inspired by the resonance phenomenon, where similar objects resonate to a given response function. The key analytic result of our work is the data terrain , which shows properties of the dataset to enable effective and efficient analysis. We demonstrated our work in the context of various real-world problems. In doing so, we establish it as the tool for preprocessing data before applying computationally expensive algorithms.

In the research of shape-based 3D model retrieval, the analysis and classification of 3D model database is an important topic for improving the retrieval performance. However, it encounters difficulties due to lack of valuable prior knowledge and the semantic gaps exist in 3D model retrieval. The paper proposes a new auto-stopped hierarchical clustering algorithm overcome these problems, which combines outlier detection with clustering. The Princeton Shape Benchmark along with 2 data sets from UCI is employed to evaluate the performance of the algorithm. And the new algorithm outperforms other auto-stopped algorithms and obtains better classification of 3D model database.

When the data consists of a set of objects described by a set of variables, we have recently proposed a new mixture model which takes into account the block clustering problem on the both sets and have developed the block CEM algorithm. In this paper, we embed the block clustering problem of contingency table in the mixture approach. In using a Poisson model and adopting the classification maximum likelihood principle we perform an adapted version of block CEM. We evaluate its performance and compare it to a simple use of CEM applied on the both sets separately. We present detailed experimental results on simulated data and we show the interest of this new algorithm.

Imbalanced data learning has recently begun to receive much attention from research and industrial communities as traditional machine learners no longer give satisfactory results. Solutions to the problem generally attempt to adapt standard learners to the imbalanced data setting. Basically, higher weights are assigned to small class examples to avoid their being overshadowed by the large class ones. The difficulty determining a reasonable weight for each example remains. In this work, we propose a scheme to weight examples of the small class based solely on local data distributions. The approach is for categorical data, and a rule learning algorithm is constructed taking the weighting scheme into account. Empirical evaluations prove the advantages of this approach.

Frequent Itemsets mining is well explored for various data types, and its computational complexity is well understood. There are methods to deal effectively with computational problems. This paper shows another approach to further performance enhancements of frequent items sets computation. We have made a series of observations that led us to inventing data pre-processing methods such that the final step of the Partition algorithm, where a combination of all local candidate sets must be processed, is executed on substantially smaller input data. The paper shows results from several experiments that confirmed our general and formally presented observations.

We present a browser extension to dynamically learn to filter unwanted Uniform Resource Locators (such as advertisements or flashy images) based on minimal user feedback. Our extension builds upon one of the top ten of Mozilla firefox plug-ins which filters URLs without learning capabilities. We apply a weighted majority-type learning algorithm working on regular expressions. Experimental results confirm that the accuracy of the predictions converges quickly to very high levels, with other key parameters: recall, specificity and precision.

Bi-clustering is a promising conceptual clustering approach. Within categorical data, it provides a collection of (possibly overlapping) bi-clusters, i.e., linked clusters for both objects and attribute-value pairs. We propose a generic framework for bi-clustering which enables to compute a bi-partition from collections of local patterns which capture locally strong associations between objects and properties. To validate this framework, we have studied in details the instance CDK-Means . It is a K-Means -like clustering on collections of formal concepts, i.e., connected closed sets on both dimensions. It enables to build bi-partitions with a user control on overlapping between bi-clusters. We provide an experimental validation on many benchmark datasets and discuss the interestingness of the computed bi-partitions.

Collaborative filtering (CF) systems are widely used by E-commerce sites to provide predictions using existing databases comprised of ratings recorded from groups of people evaluating various items, sometimes, however, such systems’ ratings are split among different parties. To provide better filtering services, such parties may wish to share their data. However, due to privacy concerns, data owners do not want to disclose data. This paper presents a privacy-preserving protocol for CF grounded on vertically partitioned data. We conducted various experiments to evaluate the overall performance of our scheme.

Sequential pattern mining has been an emerging problem in data mining. In this paper, we propose a new algorithm for mining frequent sequences. It processes only one scan of the database thanks to an indexed structure associated to a bit map representation. Thus, it allows a fast data access and a compact storage in main memory. The experimental results show the efficiency of our method compared to existing algorithms. It has been tested on synthetic data and on real data containing sequences of activities of a urban population time-use survey.