1,721,023 research outputs found
A double imputation method for Data Fusion
No full textData Fusion consists of merging information coming from two different surveys. The first is “donor survey” while the second is “receptor survey”. The aim is to complete the receptor matrix exploiting information acquired from the donor matrix. The proposed method allows to impute the missing information into the second survey through a mix of the two different methodologies proposed in literature: Explicit model-based estimation and Implicit models for imputation
A new approach to Data Fusion through Constrained Principal Component Analysis
No full textData Fusion consists of merging information coming from two different surveys. The first is called “reference” or “donor survey” while the second is called “punctual” or “receptor survey”. The aim is to complete the receptor matrix exploiting information acquired from the donor matrix. The two independent surveys have a block of common variables used as a bridge between them. In this work a Data Fusion methodology based on the Constrained Principal Component Analysis (CPCA) technique is presented. The proposed method allows to impute the missing information into the second survey taking into account knowledge about non-symmetric relationship structure among variables
Data fusion: un approccio non simmetrico al file grafting
No full textGli argomenti trattati nella tesi costituiscono quindi una trattazione organica e originale sulla problematica del file grafting nell’ambito del Data Fusion.
Nello specifico il lavoro di tesi è così articolato.
Il primo capitolo ha lo scopo di introdurre e presentare nel dettaglio il Data Fusion in tutte le sue fasi metodologiche, partendo dalla definizione del fusion e specificando i modelli e i metodi con cui vengono effettuate le fusioni.
Il secondo capitolo definisce, approfondisce e analizza sinteticamente la tecnica dell’ACPR. Nello stesso capitolo espone nel dettaglio l’utilizzo dell’ACPR per la soluzione del problema della similarità tra soggetti, e i vari passaggi che consentono l’utilizzo di una tecnica non simmetrica nel file grafting.
Il terzo capitolo pone l’attenzione sulle condizioni che consentono di effettuare il file grafting. In particolare, con la proposta di un test per la verifica della non significatività delle differenze tra le strutture delle variabili in comune alle due indagini.
Il quarto capitolo propone un nuovo algoritmo per il file grafting, provando a dare una nuova soluzione al problema della fusione dei dati. Il modello proposto denominato Grafting non simmetrico, è concepito con lo scopo di evidenziare similitudini tra i soggetti delle due indagini, alla luce della relazione, verificata esistente, tra i due gruppi di variabili che compongono l’indagine di riferimento
I fattori strutturali dell'utilizzo del lavoro non-standard in Italia: un'analisi multivariata delle imprese
No full text
The Effect of Disadvantages on Life Satisfaction
No full textSubjective well-being is an important aspect of people’s quality of life and life satisfaction is one of the key variables to measure it. However for policies to be effective in maximising subjective well-being it is of paramount importance to study the factors that influence it. This paper use microdata from the 2013 wave of Multipurpose survey on Everyday life aspects which is a large scale cross sectional survey carried out by Istat annually since 1993 to study the effect of disadvantages on people’s life satisfaction. Trees can be used for exploratory analysis in order to understand the relationships between the target variable and the predictors. We use a regression tree technique to show the effect of disadvantages on life satisfaction. The technique shows that in general the higher the level of disadvantages, the lower the life satisfaction but different combination of disadvantages have different effects on life satisfaction. The regression tree used allows to measure the importance of each factor
Modeling university student satisfaction: the case of the humanities and social studies degree programs
No full textThis paper investigates the problem of modeling students’ satisfaction ratings of various aspects of academic teaching in six humanities departments at University of Naples Federico II. In particular, we propose a strategy for analyzing data from the annual survey used to collects feedback from students across the university. The statistical procedure for this data analysis consists of two steps. First, the random forest method is fitted to the data to identify important predictors of student global satisfaction. Second, the probability distribution of student satisfaction ratings is estimated by fitting a mixture distribution with varying parameters (denoted the CUB model). The random forest methods shows that students’ interest in the course topics, together with the course objectives and teaching tools, are the main determinants of student satisfaction. Inclusion of these covariates in the CUB models confirms their dominant role in differentiating students’ evaluations of degree courses
- …
