Notice bibliographique
Résumé
After years of development, and especially since the introduction of some powerful algorithmic approaches including those for link, anchor, title analysis, and so on, web search has become amazingly powerful in its ability to discover and exploit nearly any kind of information within the billions of pages that comprise the Web. However, as powerful and large as current web search engines are, they are still limited in their ability to always deliver key services to their users especially when there is a considerable number of users with different search intentions and needs. Consequently, there is a recent merge of interest and efforts within both industry and academia in developing novel search applications to satisfy different needs. This is especially true for those application domains that have been not fully considered by traditional web search engines. The aim of the present thesis is to propose different ways of extending or using classical link analysis methods, that have been extensively applied in many classical web mining contexts, for several novel search applications which are appearing as a response to recent trends in web search. We consider different semantics related to each application domain and propose different ways of combining these semantic features with classical link analysis to enhance the corresponding web mining task. In particular, we study both analytically and empirically web communities, personalized search, geographically-oriented search, and meta-search emphasizing their retrieval aspect. While most techniques used by traditional information retrieval (IR) are mainly focused on content analysis, the modern WebIR has clearly shown that almost any web mining retrieval task can be enhanced through the exploitation of other features that are inherent in the Web (e.g. anchor text) beyond simple content. In this thesis, we again validate such hypothesis (in most cases through a set of experiments) verifying that the exploitation of features particular to each corresponding web mining task can enhance search effectiveness. For web communities, we propose a model for their extraction and the ranking of their members based on the Random Field Ising Model (RFIM), capable of representing both semantic and linkage information of web pages through node and edge weights. For personalized search, we propose an analytical model for personalized search unifying four critical aspects of the problem, namely link structure, document content, user query, and user preference. For geographically-oriented search, we study how geographic semantics can be combined with classical link analysis to improve geographically-oriented search especially in crawling and ranking aspects. For metasearch, we recast the metasearch problem as a ranking problem on a bipartite graph, constructed for a given query, and we propose a solution based on classical link analysis.
Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.
Comment cette classification a été obtenuedéplier
Prédiction machine sur la base complète
Imitation des enseignantsNi prévalence calibrée, ni vérité terrain. Validation humaine à venir. Le volet Gemma est une étiquette directe du modèle pour chaque travail de la base, lue sur la notice réduite au titre. Le volet Codex est un classifieur appris des 10 348 étiquettes directes de Codex et calibré sur les taux pondérés de l'échantillon; les champs sans appui suffisant ne portent aucun appel Codex. Le mode candidate est l'union des deux volets; le consensus est leur intersection. Ces sorties portent le statut machine_predicted_unvalidated et ne sont pas des étiquettes humaines.
Scores du classifieur distillé par catégorie (deux têtes)
| Catégorie | Codex | Gemma |
|---|---|---|
| Métarecherche | 0,004 | 0,019 |
| Méta-épidémiologie (sens strict) | 0,001 | 0,001 |
| Méta-épidémiologie (sens large) | 0,001 | 0,002 |
| Bibliométrie | 0,008 | 0,006 |
| Études des sciences et des technologies | 0,001 | 0,001 |
| Communication savante | 0,003 | 0,005 |
| Science ouverte | 0,002 | 0,003 |
| Intégrité de la recherche | 0,001 | 0,002 |
| Charge utile insuffisante (le modèle a refusé de juger) | 0,003 | 0,002 |
Scores machine (provisoires)
Les deux têtes enseignantes du modèle étudiant, lues sur ce travail. Un score ordonne la base pour la relecture; il n'affirme jamais une catégorie, et le statut de validation accompagne chaque rangée tel quel.
Scores de référence d'un modèle non mature (critères de maturité non atteints, 7 itérations). Un score ordonne; il n'affirme jamais une catégorie.
score_only:v0-immature-baseline · tel quel depuis la passe de notation : score_only signifie que le nombre peut ordonner les travaux, et qu'aucune étiquette de catégorie n'en découleClassification
machine, non validéePrédiction automatique; un appel candidat d’une seule source (Gemma direct ou Codex distillé), pas un consensus.
Le détail, modèle par modèle et score par score, se trouve en fin de page sous « Comment cette classification a été obtenue ».