1,721,058 research outputs found
Wine authenticity assessed via trimming
No full textAn authentic food is one that is what it claims to be. Consumers and
food processors need to be assured they receive exactly the specific product they pay
for. To ascertain varietal genuinity and distinguish doctored food, in this paper we
propose to employ a robust mixture estimation method. It has been shown to be a
valid tool for food authenticity studies, when applied to food data with unobserved
heterogeneity, to classify genuine wines and identify low proportions of observations
with different origins. Our methodology models the data as arising from a mixture of
Gaussian factors and employ a threshold on the multivariate density to bring apart the
less plausible data under the fitted model. Simulation results assess the effectiveness
of the proposed approach and yield very good misclassification rates when compared
to analogous methods
Monitoring tools for robust estimation of cluster weighted models = Strumenti di monitoring per la stima robusta del modello Cluster Weighted
Get PDFNella stima robusta di un cluster weighted model, lo statistico deve fare
molte scelte: specificare la forma dei cluster nelle variabili esplicative, assumere
(o meno) varianza uguale per gli errori nelle linee di regressione e impostare i va-
lori degli iper-parametri per la stima robusta, per evitare la distorsione generata
da valori anomali e contaminazione. L’iper-parametro pi`u delicato da specificare
`e la percentuale di trimming, ovvero la quantit`a di dati da escludere nella stima
per garantirne l’affidabilit`a. In questo lavoro introduciamo specifici strumenti dia-
gnostici per aiutare il professionista, o lo scienziato che ha bisogno di classificare
i dati, a compiere una scelta ragionata a riguardo di tale iper-parametro, anche in
base ad una prima esplorazione dello spazio delle soluzioni.In a robust approach to model fitting for the cluster weighted model, many choices are to be made by the statistician: specifying the shape of the clusters in the explanatory variables, assuming (or not) equal variance for the errors in the re- gression lines, and setting hyper-parameter values for the robust estimation to be protected from outliers and contamination. The most delicate hyper-parameter to specify is perhaps the percentage of trimming, or the amount of data to be excluded from the estimate, to ensure reliable inference. In this work we introduce diagnos- tic tools to help the professional, or the scientist who needs to group the data, to make an educated choice about this hyper-parameter, after a first exploration of the resulting model space
Predicting and improving smart mobility: a robust model-based approach to the BikeMi BSS = Prevedere e migliorare la mobilita smart: un approccio robusto di classificazione applicato a BikeMi
Get PDFI sistemi di Bike Sharing giocano un ruolo centrale nella mobilita sosteni- ` bile, uno dei sei pilastri che indentificano una Smart City. Motivati da un set di dati disponibile online, questo lavoro presenta l’utilizzo di due modelli di classificazione robusta per prevedere il manifestarsi di situazioni in cui una bike station sia piena e/o vuota, cos`ı creando perdita di domanda ed insoddisfazione nei clienti. Esperimenti di classificazione sulle stazioni BikeMi nel centro di Milano evidenziano l’efficacia dei metodi proposti.Bike Sharing Systems play a central role in what is identified to be one of the six pillars of a Smart City: smart mobility. Motivated by a freely available dataset, we discuss the employment of two robust model-based classifiers for pre- dicting the occurrence of situations in which a bike station is either empty or full, thus possibly creating demand loss and customer dissatisfaction. Experiments on BikeMi stations located in the central area of Milan are provided to underline the benefits of the proposed methods
Group-wise penalized estimation schemes in model-based clustering = Strategie di stima penalizzata a livello di gruppo nel clustering basato su modello
Get PDFLa sovra-parametrizzazione dei modelli di mistura Gaussiani, che rapp-
resentano un approccio probabilistico al clustering, mette a rischio la loro utilit`a in
dimensioni elevate. Per questo motivo sono state proposte strategie di stima penaliz-
zate che permettono di gestire matrici di precisioni di grandi dimensioni, sfruttando
il legame tra modelli grafici Gaussiani e modelli mistura. Questi metodi, assumendo
sparsit`a simile tra tutte le componenti, falliscono quando la struttura di dipendenza
varia di gruppo in gruppo. La nostra proposta, penalizzando una trasformazione
delle matrici di precisione differente per ogni componente, gestisce situazioni in
cui il numero di connessioni tra le variabili `e diverso tra i gruppi. La validit`a del
metodo `e evidenziata grazie ad un’applicazione a dati realiGaussian mixture models provide a probabilistically sound clustering approach.
However, their tendency to be over-parameterized endangers their utility
in high dimensions. To induce sparsity, penalized model-based clustering strategies
have been explored. Some of these approaches, exploiting the link between
Gaussian graphical models and mixtures, allow to handle large precision matrices,
encoding variables relationships. By assuming similar components sparsity levels,
these methods fall short when the dependence structures are group-dependent. Our
proposal, by penalizing group-specific transformations of the precision matrices, automatically
handles situations where under or over-connectivity between variables
is witnessed. The performances of the method are shown via a real data experimen
Position and orientation in space of bones during movement: anatomical frame definition and determination
No full textThis paper deals with methodological problems related to the reconstruction of the position and orientation of the human pelvis and the lower limb bones in space during the execution of locomotion and physical exercises using a stereophotogrammetric system. The intention is to produce a means of quantitative description of joint kinematics and dynamics for both research and application. Anatomical landmarks and bone-embedded anatomical reference systems are defined. A contribution is given to definition of variables and relevant terminology. The concept of anatomical landmark calibration is introduced and relevant experimental approaches presented. The problem of data sharing is also addressed. This material is submitted to the scientific community for consideration as a basis for standardization
Going Beyond Counting First Authors in Author Co-citation Analysis
Get PDFThe present study examines one of the fundamental aspects of author co-citation analysis (ACA) - the way co-citation
counts are defined. Co-citation counting provides the data on which all subsequent statistical analyses and mappings
are based, and we compare ACA results based on two different types of co-citation counting - the traditional type that
only counts the first one among a cited work's authors on the one hand and a non-traditional type that takes into
account the first 5 authors of a cited work on the other hand. Results indicate that the picture produced through this non-traditional author co-citation counting contains more coherent author groups and is therefore considerably clearer. However, this picture represents fewer specialties in the research field being studied than that produced through the traditional first-author co-citation counting when the same number of top-ranked authors is selected and analyzed. Reasons for these effects are discussed
- …
