Abstract-Recommender Systems are used to mitigate the information overload problem in different domains by providing personalized recommendations for particular users based on their implicit and explicit preferences. However, Item-based Collaborative Filtering (CF) techniques, as the most popular techniques of recommender systems, suffer from sparsity and new item limitations which result in producing inaccurate recommendations. The use of items' semantic information besides the inclusion of multi-criteria ratings can successfully alleviate such problems and generate more accurate recommendations. This paper proposes an Item-based MultiCriteria Collaborative Filtering algorithm that integrates the items' semantic information and multi-criteria ratings of items to lessen known limitations of the item-based CF techniques. According to the experimental results, the proposed algorithm prove to be very effective in terms of dealing with both of the sparsity and new item problems and therefore produce more accurate recommendations when compared to standard itembased CF techniques.

This chapter illustrates the use of the combinatorial optimization models presented in Chapter 19 for the Feature Set selection and Gene Ordering problems to find genetic signatures for diseases using micro-array data. We demonstrate the quality of this approach by using a microarray dataset from a mouse model of Parkinson's disease. The results are accompanied by a description of the currently known molecular functions and biological processes of the genes in the signatures.

Abstract. Three different methodologies have been applied to microarray data from brains of Alzheimer diagnosed patients and healthy patients taken as control. A clear pattern of differential gene expression results which can be regarded as a molecular signature of the disease. The results show the complementarity of the different methodologies, suggesting that a unified approach may help to uncover complex genetic risk factors not currently discovered with a single method. We also compare the set of genes in these differential patterns with those already reported in the literature.

This work presents a study on the genetic profile of the left and right hemispheres of the brain of a mouse model of Parkinson's disease (PO). The goal is to characterize, in a genetic basis, PO as a disease that affects these two brain regions in different ways. Using the same wholegenome micro array expression data introduced by Brown et al. (2002) [1], we could find significant differences in the expression of some key genes, well-known to be involved in the mechanisms of dopamine production control and PD. The problem of selecting such genes was modeled as the MIN (a,/3)-FEATURE SET problem [2]; a similar approach to that employed previously to find biomarkers for different types of cancer using gene expression microarray data [3]. The Feature Selection method produced a series of genetic signatures for PO, with distinct expression profiles in the Parkinson's model and control mice experiments. In addition, a close examination of the genes composing those signatures shows that many of them belong to genetic pathways or have ontology annotations considered to be involved in the onset and development of PD. Such elements could provide new clues on which mechanisms are implicated in hemisphere differentiation in PD.

Gene expression data often contain missing expression values. For the purpose of conducting an effective clustering analysis and since many algorithms for gene expression data analysis require a complete matrix of gene array values, choosing the most effective missing value estimation method is necessary. In this paper, the most commonly used imputation methods from literature are critically reviewed and analyzed to explain the proper use, weakness and point the observations on each published method. From the conducted analysis, we conclude that the Local Least Square (LLS) and Support Vector Regression (SVR) algorithms have achieved the best performances. SVR can be considered as a complement algorithm for LLS especially when applied to noisy data. However, both algorithms suffer from some deficiencies presented in choosing the value of Number of Selected Genes (K) and the appropriate kernel function. To overcome these drawbacks, the need for new method that automatically chooses the parameters of the function and it also has an appropriate computational complexity is imperative.