Patrick Marcel

A personalization framework for OLAP queries

OLAP users heavily rely on visualization of query answers for their interactive analysis of massive amounts of data. Very often, these answers cannot be visualized entirely and the user has to navigate through them to find relevant facts.In this paper, we propose a framework for personalizing OLAP queries. In this framework, the user is asked to give his (her) preferences and a visualization constraint, that can be for instance the limitations imposed by the device used to display the answer to a query. Given this, for each query, our method computes the part of the answer that respects both the user preferences and the visualization constraint. In addition, a personalized structure for the visualization is proposed.

Recommending Multidimensional Queries

Interactive analysis of datacube, in which a user navigates a cube by launching a sequence of queries is often tedious since the user may have no idea of what the forthcoming query should be in his current analysis. To better support this process we propose in this paper to apply a Collaborative Work approach that leverages former explorations of the cube to recommend OLAP queries. The system that we have developed adapts Approximate String Matching, a technique popular in Information Retrieval, to match the current analysis with the former explorations and help suggesting a query to the user. Our approach has been implemented with the open source Mondrian OLAP server to recommend MDX queries and we have carried out some preliminary experiments that show its efficiency for generating effective query recommendations.

A framework for recommending OLAP queries

An OLAP analysis session can be defined as an interactive session during which a user launches queries to navigate within a cube. Very often choosing which part of the cube to navigate further, and thus designing the forthcoming query, is a difficult task. In this paper, we propose to use what the OLAP users did during their former exploration of the cube as a basis for recommending OLAP queries to the user. We present a generic framework that allows to recommend OLAP queries based on the OLAP server query log. This framework is generic in the sense that changing its parameters changes the way the recommendations are computed. We show how to use this framework for recommending simple MDX queries and we provide some experimental results to validate our approach.

Mining preferences from OLAP query logs for proactive personalization

The goal of personalization is to deliver information that is relevant to an individual or a group of individuals in the most appropriate format and layout. In the OLAP context personalization is quite beneficial, because queries can be very complex and they may return huge amounts of data. Aimed at making the users experience with OLAP as plain as possible, in this paper we propose a proactive approach that couples an MDX-based language for expressing OLAP preferences to a mining technique for automatically deriving preferences. First, the log of past MDX queries issued by that user is mined to extract a set of association rules that relate sets of frequent query fragments; then, given a specific query, a subset of pertinent and effective rules is selected; finally, the selected rules are translated into a preference that is used to annotate the users query. A set of experimental results proves the effectiveness and efficiency of our approach.

Similarity measures for OLAP sessions

OLAP queries are not normally formulated in isolation, but in the form of sequences called OLAP sessions. Recognizing that two OLAP sessions are similar would be useful for different applications, such as query recommendation and personalization; however, the problem of measuring OLAP session similarity has not been studied so far. In this paper, we aim at filling this gap. First, we propose a set of similarity criteria derived from a user study conducted with a set of OLAP practitioners and researchers. Then, we propose a function for estimating the similarity between OLAP queries based on three components: the query group-by set, its selection predicate, and the measures required in output. To assess the similarity of OLAP sessions, we investigate the feasibility of extending four popular methods for measuring similarity, namely the Levenshtein distance, the Dice coefficient, the tf–idf weight, and the Smith–Waterman algorithm. Finally, we experimentally compare these four extensions to show that the Smith–Waterman extension is the one that best captures the users’ criteria for session similarity.

A collaborative filtering approach for recommending OLAP sessions

While OLAP has a key role in supporting effective exploration of multidimensional cubes, the huge number of aggregations and selections that can be operated on data may make the user experience disorientating. To address this issue, in the paper we propose a recommendation approach stemming from collaborative filtering. We claim that the whole sequence of queries belonging to an OLAP session is valuable because it gives the user a compound and synergic view of data; for this reason, our goal is not to recommend single OLAP queries but OLAP sessions. Like other collaborative approaches, ours features three phases: (i) search the log for sessions that bear some similarity with the one currently being issued by the user; (ii) extract the most relevant subsessions; and (iii) adapt the top-ranked subsession to the current users session. However, it is the first that treats sessions as first-class citizens, using new techniques for comparing sessions, finding meaningful recommendation candidates, and adapting them to the current session. After describing our approach, we discuss the results of a large set of effectiveness and efficiency tests based on different measures of recommendation quality. We propose an approach for recommending OLAP sessions.Sessions similar to the current one are searched in the log, ranked, and adapted.The properties of recommendations are relevance, foresight, novelty, and suitabilityThe approach is validated using different measures of recommendation quality

Predicting Your Next OLAP Query Based on Recent Analytical Sessions

In Business Intelligence systems, users interact with data warehouses by formulating OLAP queries aimed at exploring multidimensional data cubes. Being able to predict the most likely next queries would provide a way to recommend interesting queries to users on the one hand, and could improve the efficiency of OLAP sessions on the other. In particular, query recommendation would proactively guide users in data exploration and improve the quality of their interactive experience. In this paper, we propose a framework to predict the most likely next query and recommend this to the user. Our framework relies on a probabilistic user behavior model built by analyzing previous OLAP sessions and exploiting a query similarity metric. To gain insight in the recommendation precision and on what parameters it depends, we evaluate our approach using different quality assessments.

Describing analytical sessions using a multidimensional algebra

Recent efforts to support analytical tasks over relational sources have pointed out the necessity to come up with flexible, powerful means for analyzing the issued queries and exploit them in decisionoriented processes (such as query recommendation or physical tuning). Issued queries should be decomposed, stored and manipulated in a dedicated subsystem. With this aim, we present a novel approach for representing SQL analytical queries in terms of a multidimensional algebra, which better characterizes the analytical efforts of the user. In this paper we discuss how an SQL query can be formulated as a multidimensional algebraic characterization. Then, we discuss how to normalize them in order to bridge (i.e., collapse) several SQL queries into a single characterization (representing the analytical session), according to their logical connections.

A framework for pattern-based global models

Discovering global models on a dataset (e.g., classifiers, clusterings, summaries) has attracted a lot of attention and many approaches can be found in the literature. However no framework has been proposed yet for describing and comparing these approaches in a uniform manner. In this paper we propose such a framework for pattern-based modeling approaches, i.e., approaches that use local patterns to construct a global model. This framework includes a generic algorithm (IGMA) for constructing a global model. We show that the framework allows to describe in an as declarative as possible way various different global model construction methods.

Towards intensional answers to OLAP queries for analytical sessions

One of the problems in analyzing large multidimensional databases through OLAP sessions is that decision makers can be overwhelmed by the size of query answers, while they need a concise summary of data. Intensional query answering can help by providing a concise description of extensional answers (i.e., the sets of retrieved facts), generally relying on knowledge like integrity constraints, taxonomies, or patterns discovered from data. This paper proposes a framework for computing an intensional answer to an OLAP query by leveraging on the previous queries in the current session. Such intensional answer is concise and semantically rich, and allows the size of the extensional answers returned to be reduced, so as to achieve an effective trade-off between conciseness and informational content. After describing the general framework, we propose a specific instantiation that relies on previous contributions in cube modeling and intensional query answering.