Héctor D. Menéndez

Extracting Collective Trends from Twitter Using Social-Based Data Mining

Social Networks have become an important environment for Collective Trends extraction. The interactions amongst users provide information of their preferences and relationships. This information can be used to measure the influence of ideas, or opinions, and how they are spread within the Network. Currently, one of the most relevant and popular Social Network is Twitter. This Social Network was created to share comments and opinions. The information provided by users is specially useful in different fields and research areas such as marketing. This data is presented as short text strings containing different ideas expressed by real people. With this representation, different Data Mining and Text Mining techniques (such as classification and clustering) might be used for knowledge extraction trying to distinguish the meaning of the opinions. This work is focused on the analysis about how these techniques can interpret these opinions within the Social Network using information related to IKEA® company.

Extracting behavioural models from 2010 FIFA world cup

The FIFA World Cup™ is the most profitable worldwide event. The FIFA publishes global statistics of this competition which provide global data about the players and teams during the competition. This work is focused on the extraction of behavioural patterns for both, players and teams strategies, through the automated analysis of this dataset. The knowledge and models extracted in this work could be applied to soccer leagues or even it could be oriented to sport betting. However, the main contribution is related to the study on several automatic knowledge extraction techniques, such as clustering methods, and how these techniques can be used to obtain useful behavioural models from a global statistics dataset. The information provided by the clustering algorithms shows similar properties which have been combined to define the models, making the human interpretation of these statistics easier. Finally, the most successful teams strategies have been analysed and compared.

A genetic graph-based clustering algorithm

The interest in the analysis and study of clustering techniques have grown since the introduction of new algorithms based on the continuity of the data, where problems related to image segmentation and tracking, amongst others, makes difficult the correct classification of data into their appropriate groups, or clusters. Some new techniques, such as Spectral Clustering (SC), uses graph theory to generate the clusters through the spectrum of the graph created by a similarity function applied to the elements of the database. The approach taken by SC allows to handle the problem of data continuity though the graph representation. Based on this idea, this study uses genetic algorithms to select the groups using the same similarity graph built by the Spectral Clustering method. The main contribution is to create a new algorithm which improves the robustness of the Spectral Clustering algorithm reducing the dependency of the similarity metric parameters that currently affects to the performance of SC approaches. This algorithm, named Genetic Graph-based Clustering (GGC), has been tested with different synthetic and real-world datasets, the experimental results have been compared against classical clustering algorithms like K-Means, EM and SC.

Using the clustering coefficient to guide a genetic-based communities finding algorithm

Finding communities in networks is a hot topic in several research areas like social network, graph theory or sociology among others. This work considers the community finding problem as a clustering problem where an evolutionary approach can provide a new method to find overlapping and stable communities in a graph. We apply some clustering concepts to search for new solutions that use new simple fitness functions which combine network properties with the clustering coefficient of the graph. Finally, our approach has been applied to the Eurovision contest dataset, a well-known social-based data network, to show how communities can be found using our method.

A Co-Evolutionary Multi-Objective approach for a K-adaptive graph-based clustering algorithm

Clustering is a field of Data Mining that deals with the problem of extract knowledge from data blindly. Basically, clustering identifies similar data in a dataset and groups them in sets named clusters. The high number of clustering practical applications has made it a fertile research topic with several approaches. One recent method that is gaining popularity in the research community is Spectral Clustering (SC). It is a clustering method that builds a similarity graph and applies spectral analysis to preserve the data continuity in the cluster. This work presents a new algorithm inspired by SC algorithm, the Co-Evolutionary Multi-Objective Genetic Graph-based Clustering (CEMOG) algorithm, which is based on the Multi-Objective Genetic Graph-based Clustering (MOGGC) algorithm and extends it by introducing an adaptative number of clusters. CEMOG takes an island-model approach where each island keeps a population of candidate solutions for ki clusters. Individuals in the islands can migrate to encourage genetic diversity and the propagation of individuals around promising search regions. This new approach shows its competitive performance, compared to several classical clustering algorithms (EM, SC and K-means), through a set of experiments involving synthetic and real datasets.

MACOC: A Medoid-Based ACO Clustering Algorithm

The application of ACO-based algorithms in data mining is growing over the last few years and several supervised and unsupervised learning algorithms have been developed using this bio-inspired approach. Most recent works concerning unsupervised learning have been focused on clustering, showing great potential of ACO-based techniques. This work presents an ACO-based clustering algorithm inspired by the ACO Clustering (ACOC) algorithm. The proposed approach restructures ACOC from a centroid-based technique to a medoid-based technique, where the properties of the search space are not necessarily known. Instead, it only relies on the information about the distances amongst data. The new algorithm, called MACOC, has been compared against well-known algorithms (K-means and Partition Around Medoids) and with ACOC. The experiments measure the accuracy of the algorithm for both synthetic datasets and real-world datasets extracted from the UCI Machine Learning Repository.

Combining graph connectivity and genetic clustering to improve biomedical summarization

Automatic summarization is emerging as a feasible instrument to help biomedical researchers to access online literature and face information overload. The Natural Language Processing community is actively working toward the development of effective summarization applications; however, automatic summaries are sometimes less informative than the user needs. In this work, our aim is to improve a summarization graph-based process combining genetic clustering with graph connectivity information. In this way, while genetic clustering allows us to identify the different topics that are dealt with in a document, connectivity information (in particular, degree centrality) allows us to asses and exploit the relevance of the different topics. Our automatic summaries are compared with others produced by commercial and research applications, to demonstrate the appropriateness of using this combination of techniques for automatic summarization.

A genetic graph-based clustering approach to biomedical summarization

Summarization techniques have become increasingly important over the last few years, specially in biomedical research, where information overload is major problem. Researchers of this area need a shorter version of the texts which contains all the important information while discarding irrelevant one. There are several applications which deal with this problem, however, these applications are sometimes less informative than the user needs. This work deals with this problem trying to improve a summarization graph-based process using genetic clustering techniques. Our automatic summaries are compared to those produced by several commercial and research summarizers, and demonstrate the appropriateness of using genetic techniques in automatic summarization.

Combining Time Series and Clustering to Extract Gamer Profile Evolution

Video-games industry is specially focused on user entertainment. It is really important for these companies to develop interactive and usable games in order to satisfy their client preferences. The main problem for the game developers is to get information about the user behaviour during the game-play. This information is important, specially nowadays, because gamers can buy new extra levels, or new games, interactively using their own consoles. Developers can use the gamer profile extracted from the game-play to create new levels, adapt the game to different user, recommend new video games and also match up users. This work tries to deal with this problem. Here, we present a new game, called “Dream”, whose philosophy is based on the information extraction process focused on the player game-play profile and its evolution. We also present a methodology based on time series clustering to group users according to their profile evolution. This methodology has been tested with real users which have played Dream during several rounds.

Features selection from high-dimensional web data using clustering analysis

The features selection methodologies have become an important field of the data preprocessing techniques. These methods are applied to reduced the dimension of the attributes of different datasets to simplify their analysis. Some of the classical techniques used are wrapper approaches, heuristic functions and filters. The main problem of these approaches is that they usually are black box and computationally expensive algorithms. This work presents a new straightforward strategy to reduce the dimension of the attributes. This new methodology cares about the variables distribution and has been oriented to clustering analysis. It provides an easier human interpretation of the attributes selection strategy and the resulting clusters. Finally, this new approach has been experimentally tested using the FIFA World Cup web dataset, a well-known social-based statistical data with a high number of variables, to show how the features selection strategy find the most relevant variables.