Sequential patterns mining is an important pattern discovery technique used to identify frequently observed sequential occurrence of items across ordered transactions over time. It has been intensively studied and there exists a great diversity of algorithms. However, there is a major problem associated with the conventional sequential patterns mining in that patterns derived are often large and not very easy to understand or use. In addition, more complex relations among events are often hidden behind sequences. A novel model for sequential patterns called Sequential Patterns Graph (SPG) is proposed. The construction algorithm of SPG is presented with experimental results to substantiate the concept. The thesis then sets out to define some new structural patterns such as concurrent branch patterns, exclusive patterns and iterative patterns which are generally hidden behind sequential patterns. Finally, an integrative framework, named Post Sequential Patterns Mining (PSPM), which is based on sequential patterns mining, is also proposed for the discovery and visualisation of structural patterns. This thesis is intended to prove that discrete sequential patterns derived from traditional sequential patterns mining can be modelled graphically using SPG. It is concluded from experiments and theoretical studies that SPG is not only a minimal representation of sequential patterns mining, but it also represents the interrelation among patterns and establishes further the foundation for mining structural knowledge (i.e. concurrent branch patterns, exclusive patterns and iterative patterns). from experiments conducted on both synthetic and real datasets, it is shown that Concurrent Branch Patterns (CBP) mining is an effective and efficient mining algorithm suitable for concurrent branch patterns.

Cette thèse se place dans le cadre de services web en dépassant leur description pour considérer leur structuration en réseaux (réseaux d'interaction et réseaux de similitude). Nous proposons des méthodes basées sur les motifs, la modélisation probabiliste et l'analyse des concepts formels, pour améliorer la qualité des services découverts. Trois contributions sont alors proposées: découverte de services diversifiés, recommandation de services et cohérence des communautés de services détectées. Nous structurons d'abord les services sous forme de réseaux. Afin de diversifier les résultats de la découverte, nous proposons une méthode probabiliste qui se base à la fois sur la pertinence, la diversité et la densité des services. Dans le cas de requêtes complexes, nous exploitons le réseau d'interaction de services construit et la notion de diversité dans les graphes pour identifier les services web qui sont susceptibles d'être composables. Nous proposons également un système de recommandation hybride basé sur le contenu et le filtrage collaboratif. L'originalité de la méthode proposée vient de la combinaison des modèles thématiques et les motifs fréquents pour capturer la sémantique commune maximale d'un ensemble de services. Enfin, au lieu de ne traiter que des services individuels, nous considérons aussi un ensemble de services regroupés sous forme de communautés de services pour la recommandation. Nous proposons dans ce contexte, une méthode qui combine la sémantique et la topologie dans les réseaux afin d'évaluer la qualité et la cohérence sémantique des communautés détectées, et classer également les algorithmes de détection de communautés. / As a part of this thesis, we exceed the description of web services to consider their structure as networks (i.e. similarity and interaction web service networks). We propose methods based on patterns, topic models and formal concept analysis, to improve the quality of discovered services. Three contributions are then proposed: (1) diversified services discovery, (2) services recommendation and (3) consistency of detected communities. Firstly, we propose modeling the space of web services through networks. To discover the diversified services corresponding to a given query, we propose a probabilistic method to diversify the discovery results based on relevancy, diversity and service density. In case of complex requests, it is necessary to combine multiple web services to fulfill this kind of requests. In this regard, we use the interaction web service network and the diversity notion in graphs to identify all possible services compositions. We also propose a new hybrid recommendation system based on both content and collaborative filtering. Its originality comes from the combination of probabilistic topic models and pattern mining to capture the maximal common semantic of a set of services. Finally, instead of processing individual services, we consider a set of services grouped into service communities for the recommendation. We propose in this context, a new method combining both topology and semantics to evaluate the quality and the semantic consistency of detected communities, and also rank the detection communities algorithms.