Francesco Ricci

Free University of Bozen-Bolzano
Faculty of Computer Science

Piazza Domenicani 3, I-39100 Bozen-Bolzano, Italy
Phone: 0471 016 971, fax: +39 0471 016 009
email: fricci At

A photo of Francesco Ricci

Teaching | Research Interests and Activities | Projects | Publications | CV-eng 

Yellow point Teaching and Students' Supervision I'm looking for students interested in starting a PhD on Recommender Systems. Please send me your CV.

I' m currently supervising or I have recently supervised the following PhD and Master students:

Yellow point Research Interests and Activities

I'm involved in research activities on recommender systems and intelligent information systems. Please look at Recommender Systems Handbook, which I edited with L. Rokach, B. Shapira and P. Kantor, for a extensive overview with several papers covering several research topics in this area. The new version is supposed to be published in 2015.

I'm interested in the design of innovative recommendation technologies and in the development of concrete and fully operational recommender systems. I participated to the design and development of: NutKing, DieToRecs, and MobyRek. These recommenders integrate etherogeneous user-related and product-related data sources, support cooperative query answering, and exploit a structured case representation and novel similarity metrics.
Recommender Systems Handbook

The core recommendation technology that I designed, Trip@dvice, has been reenginered in the European Tourist Destination Portal Project and it is fully exploited in the European Travel Commission web site, in the Finnish travel portal and in A spin off company,, is marketing the Trip@advice technology.

A detailed description of my recent research projects can be found here.

More in general my research interests are:

I've been recently invited to give talks or lectures in the following venues:
My current scientific activities are:
Yellow point Research Summary

Recommender systems

I'm studying hybrid (case- content- collaborative-based) methodologies for building recommender systems for complex products like travel plans. The methodology exploits idea from cooperative data base, case-based reasoning, collaborative-filtering and decision theory, to supports a user in build his own personalized itinerary. We have developed a system prototype (NutKing) that integrates data and information originating from external tourist portals exploiting an XML-based information server and data mapping techniques. (ppt presentation). I've also been involved in research on mobile recommender systems, and in particular in the application of critiques-based technologies to support on-the-move decision making (MobiRek).

Machine Learning

I've studied local similarity metrics for case-based reasoning and Nearest Neighbor classification. I've developed a learning procedure for adapting the features' weights to the input space. The goal is to achieve the same accuracy obtained with other Nearest Neighbour algorithms with less cases in the memory (data compression), so obtaining also performance improvements. I've been also studying a class of local metrics for the nearest neighbor classifier whose definition is based on statistics computed on the case base. These metrics perform comparably to the Bayes classifier based on the same probability estimates. Finally, I've also integrated local learners with error-correcting output codes. Error-correcting output codes (ECOCs) represent classes with a set of output bits, where each bit encodes a binary classification task corresponding to a unique partition of the classes. Algorithms that use ECOCs learn the function corresponding to each bit, and combine them to generate class predictions. ECOCs can reduce both variance and bias errors for multiclass classification tasks when the errors made at the output bits are not correlated.

Case Based Reasoning

I developed a software tool for the interactive exploration of a case base. In this tool well rooted ML techniques for selecting relevant features, clustering cases and forecasting unknown values have been adapted and reused for case base exploration. I've studied algorithms for case retrieval from a case base of trees labeled on both nodes and edges. They extends the best known algorithm for solving the subtree-isomorphism problem. That algorithm is based on branch and bound and on a general similarity metric between trees. Moreover I've applied case-based reasoning to recommender systems and developed case-based storage and retrieval technologies based on a P2P infrastructure.

Emergency management and case-based planning

We have developed a CBR architecture for planning that integrate constraint processing to support temporal reasoning. A prototype was developed within an Esprit project (CHARADE) whose goal was to fully support forest fire control centers, from situation assessment to resource dispatching.

Constraint Reasoning

I've studied a decision maker model, called learning automaton, exhibiting adaptive behavior in highly uncertain stochastic environments. This learning model have been used in solving constraint satisfaction problems (CSPs) by a procedure that can be viewed as hill climbing in probability space. Moreover, I've studied the application of incremental algorithm for constraint satisfaction problems. In particular I'm interested in designing incremental algorithms for bounded difference constraints networks. That type of constraint networks have been applied for solving temporal reasoning (planning and scheduling) and design problems.

Valid HTML 4.01 Transitional Last update: May 25th, 2015