Cristina Davino

Visual Mining of Association Rules

Association Rules are one of the most widespread data mining tools because they can be easily mined, even from very huge database, and they provide valuable information for many application fields such as marketing, credit scoring, business, etc. The counterpart is that a massive effort is required (due to the large number of rules usually mined) in order to make actionable the retained knowledge. In this framework vizualization tools become essential to have a deep insight into the association structures and interactive features have to be exploited for highlighting the most relevant and meaningful rules.

Visual post-analysis of association rules ☆

Abstract Association rules (AR) represent a consolidated tool in data mining applications as they are able to discover regularities in large data sets. The information mined by the rules is very often difficult to exploit because of the presence of too many associations where to detect the really relevant logical implications. In this framework, by combining methodological and graphical pruning techniques, AR post-analysis tools are proposed. The methodological techniques will ensure the statistical significance of the AR which were not pruned, while the graphical ones will provide interactive and powerful visualization tools.

Quantile composite-based path modeling

The paper aims at introducing a quantile approach in the Partial Least Squares path modeling framework. This is a well known composite-based method for the analysis of complex phenomena measurable through a network of relationships among observed and unobserved variables. The proposal intends to enhance potentialities of the Partial Least Squares path models overcoming the classical exploration of average effects. The introduction of Quantile Regression and Correlation in the estimation phases of the model allows highlighting how and if the relationships among observed and unobserved variables change according to the explored quantile of interest. The proposed method is applied to two real datasets in the customer satisfaction measurement and in the sensory analysis framework but it proves to be useful also in other applicative contexts.

Assessment and Validation in Quantile Composite-Based Path Modeling

The paper aims to introduce assessment and validation measures in Quantile Composite-based Path modeling. A quantile approach in the Partial Least Squares path modeling framework overcomes the classical exploration of average effects and highlights how and if the relationships among observed and unobserved variables change according to the explored quantile of interest. A final evaluation of the quality of the obtained results both from a descriptive (assessment) and inferential (validation) point of view is needed. The functioning of the proposed method is shown through a real data application in the area of the American Customer Satisfaction Index.

Assessing Multi-Item Scales for Subjective Measurement

In this paper a method for assessing different multi-item scales in subjective measurement is described and discussed. The method is a combination of analysis of variance models and multivariate techniques. It allows us comparisons among the scales by considering the multivariate information provided by the items. Focus is given on the way individual differences in the use of the scales may be interpreted and crossed with respondent characteristics. The approach is illustrated by analysing data from a survey on the assessment of students’ university quality of life.

Analyzing Research Potential through Redundancy Analysis: the case of the Italian University System

The paper proposes a multivariate approach to study the dependence of the scientific productivity on the human research potential in the Italian University system. In spite of the heterogeneity of the system, Redundancy Analysis is exploited to analyse the University research system as a whole. The proposed approach is embedded in an exploratory data analysis framework.

Structural Neural Networks for Modeling Customer Satisfaction

The aim of this paper is to provide a Structural Neural Network to model Customer Satisfaction in a business-to-business framework. Neural Networks are proposed as a complementary approach to PLS path modeling, one of the most widespread approaches for modeling and measuring Customer Satisfaction. The proposed Structural Neural Network allows to overcome one of the main drawbacks of Neural Networks because they are usually considered as black boxes.

Quantile Composite-Based Model: A Recent Advance in PLS-PM

The aim of the present chapter is to discuss a recent contribution in the partial least squares path modeling framework: the quantile composite-based path modeling. We introduce this recent contribution from both a methodological and an applicative point of view. The objective is to provide an exploration of the whole dependence structure and to highlight whether and how the relationships among variables (both observed and unobserved) change across quantiles. We use a real data application, measuring the equitable and sustainable well-being of Italian provinces. Partial least squares path modeling is first applied to study the relationships among variables assuming homogeneity among observations. Afterwards, a multi-group analysis is performed, assuming that a specific factor (the geographic area) causes heterogeneity in the population. Finally, the quantile approach to composite-based path modeling provides a more in-depth analysis. Some relevant results are selected and described to show that the quantile composite-based path modeling can be very useful in this real data application, as it allows us to explore territorial disparities in depth.

Neural Networks Applications in Economics: a Statistical Point of View

Nowdays neural networks (NN) are applied in the most various fields and are actually receiving a lot of attention among the researcher’s community. In this paper we will provide a review of some NN applications in economics. We distinguish the applications according to the main objectives achieved by NN in this field: prediction, classification and modeling economic theory. It is a matter of fact that NN share with statistics a lot of methodological and computational aspects [9] as well as many fields of application. In this framework we introduce a general strategy for a statistical approach to NN which allows to use NN in a statistical context taking into account typical statistical applications problems, such as the selection and coding of the variables, the sample representativeness but more the interpretation, visualization and stability of the results.