Spelling suggestions: "subject:"[een] MULTIVARIATE ANALYSIS"" "subject:"[enn] MULTIVARIATE ANALYSIS""
91 |
Statistical inference of a threshold model in extreme value analysisLee, David., 李大為. January 2012 (has links)
In many data sets, a mixture distribution formulation applies when it is
known that each observation comes from one of the underlying categories. Even
if there are no apparent categories, an implicit categorical structure may justify
a mixture distribution. This thesis concerns the modeling of extreme values in
such a setting within the peaks-over-threshold (POT) approach. Specifically,
the traditional POT modeling using the generalized Pareto distribution is augmented
in the sense that, in addition to threshold exceedances, data below the
threshold are also modeled by means of the mixture exponential distribution.
In the first part of this thesis, the conventional frequentist approach is
applied for data modeling. In view of the mixture nature of the problem,
the EM algorithm is employed for parameter estimation, where closed-form
expressions for the iterates are obtained. A simulation study is conducted to
confirm the suitability of such method, and the observation of an increase in
standard error due to the variability of the threshold is addressed. The model
is applied to two real data sets, and it is demonstrated how computation time
can be reduced through a multi-level modeling procedure. With the fitted
density, it is possible to derive many useful quantities such as return periods
and levels, value-at-risk, expected tail loss and bounds for ruin probabilities.
A likelihood ratio test is then used to justify model choice against the simpler
model where the thin-tailed distribution is homogeneous exponential.
The second part of the thesis deals with a fully Bayesian approach to the
same model. It starts with the application of the Bayesian idea to a special
case of the model where a closed-form posterior density is computed for the
threshold parameter, which serves as an introduction. This is extended to
the threshold mixture model by the use of the Metropolis-Hastings algorithm
to simulate samples from a posterior distribution known up to a normalizing
constant. The concept of depth functions is proposed in multidimensional
inference, where a natural ordering does not exist. Such methods are then
applied to real data sets. Finally, the issue of model choice is considered
through the use of posterior Bayes factor, a criterion that stems from the
posterior density. / published_or_final_version / Statistics and Actuarial Science / Master / Master of Philosophy
|
92 |
A polytomous nonlinear mixed model for item analysisShin, Seon-hi 25 July 2011 (has links)
Not available / text
|
93 |
Ανάλυση σε κύριες συνιστώσες και παραγοντική ανάλυσηΓκίτσης, Σπυρίδων 29 August 2008 (has links)
Η ανάλυση πολυμεταβλητών δεδομένων καθίσταται ιδιαίτερα δύσκολη όταν το
πλήθος των μεταβλητών, p (διάσταση των δεδομένων), είναι μεγάλο. Επίσης
δυσκολία υπάρχει στην ανάλυση, όταν οι μεταβλητές είναι υψηλά συσχετισμένες
μεταξύ τους.
Η ανάλυση κύριων συνιστωσών είναι πολυμεταβλητή στατιστική τεχνική που
ασχολείται με την δομή διασπορών – συνδιασπορών, μέσω μερικών γραμμικών
συνδυασμών των αρχικών μεταβλητών. Γενικότερα τα αντικείμενα της είναι (1) η
μείωση των δεδομένων και (2) η ανάλυση (ερμηνεία) τους.
Παρόλο που απαιτούνται p μεταβλητές για να ερμηνευτεί η συνολική
μεταβλητότητα του συστήματος, συχνά, η περισσότερη από αυτή τη μεταβλητότητα
μπορεί να ερμηνευτεί από ένα μικρό αριθμό k κύριων συνιστωσών. Αν πράγματι
συμβεί αυτό, τότε, υπάρχει (σχεδόν) τόση πληροφορία στις k συνιστώσες, όση
υπάρχει στις p αρχικές μεταβλητές. Οι k κύριες συνιστώσες μπορούν τότε να
αντικαταστήσουν τις αρχικές p μεταβλητές, και το αρχικό σύνολο δεδομένων που
αποτελείται από n μετρήσεις των p μεταβλητών, μειώνεται σε ένα σύνολο δεδομένων
που αποτελείται από n μετρήσεις των k μεταβλητών. Οι k κύριες συνιστώσες είναι
γραμμικός συνδυασμός των p αρχικών μεταβλητών, και μάλιστα είναι ασυσχέτιστες
μεταξύ τους. Έτσι, οδηγούμαστε από ένα σύνολο p συσχετισμένων μεταβλητών, σ’
ένα μικρότερο σύνολο k ασυσχέτιστων μεταβλητών.
Η μείωση αυτή των δεδομένων είναι πολύ σημαντικό γεγονός, διότι αντί να
αναλύουμε δεδομένα στο R
p
, αναλύουμε δεδομένα στο R
k
. Σε ορισμένες περιπτώσεις
το k, η νέα διάσταση, είναι 2 ή 3 και τότε έχουμε μια οπτική ιδέα, μια εικόνα των
δεδομένων.
Κλείνοντας την εισαγωγή, θα πρέπει να αναφέρουμε ότι η τεχνική κύριων
συνιστωσών δεν επιτυγχάνει πάντοτε την μείωση της διάστασης, π.χ., αυτό συμβαίνει
όταν οι αρχικές μεταβλητές είναι ασυσχέτιστες. Τότε θα πρέπει να αναζητηθούν
άλλες μέθοδοι μείωσης της διάστασης. / -
|
94 |
Quadratic forms in normal variablesScarowsky, Issie January 1973 (has links)
No description available.
|
95 |
Structural properties of convolutional codes : an algorithmic approach with applications to linear multivariable system theoryConan, Jean. January 1980 (has links)
A new approach to the analysis of the structural properties of multivariable convolutional codes over finite fields is presented. It is based on the properties of the state transition graph which can be considered as a generalization to the multivariable case of the classical Good-De-Bruijn graph associated with linear shift register sequences. The concept of a minimal graph is introduced and shown to be isomorphic to the class of all minimal encoders previously defined by Forney. Straightforward algorithms based on simple algebraic and graph manipulations are introduced to allow for the reduction of any state transition graph to a minimal form. Furthermore each stage in the reduction procedure is shown to be related to some fundamental system theoretic concept including the conditions for causal invertibility, pseudo invertibility and polynomial invertibility of a linear feedforward system. By using the concept of dual codes and introducing a straightforward algorithm for the construction of a dual encoder in minimal form which is valid on any field; a simple procedure is further devised providing for the reduction of any rational basis to a minimal polynomial form and the applications of this result to multivariable realization theory are discussed. Finally several non exhaustive applications of the above mentioned concepts to linear system theory are developed. A special emphasis is placed on the solution of the problem associated with the construction of the class of all minimal order, minimal delay pseudo inverses of any realizable linear system. Furthermore, we present a solution to the minimal partial realization problem for vectored sequences based on the use of a Berlekamp-Massey type algorithm.
|
96 |
Determination of banned sudan dyes in culinary spices through spectroscopic techniques and multivariate analysisDi Anibal, Carolina Vanesa 19 December 2011 (has links)
La presente tesis esta focalizada en el desarrollo de métodos analíticos para determinar la adulteración de especias culinarias con colorantes Sudan I, II, III y IV. Estos colorantes están prohibidos como aditivos para uso alimentario por la legislación europea ya que son carcinógenos. Las metodologías analíticas desarrolladas están basadas en el uso de técnicas espectroscópicas como UV-visible, Resonancia Magnética de protón y Raman junto con tratamiento multivariante de los datos obtenidos. En relación al análisis multivariante, como principal objetivo se planteó el establecimiento de modelos de clasificación y posteriormente se utilizaron diversas herramientas quimiométricas con el objetivo de mejorar los resultados de clasificación: análisis exploratorio de datos, métodos de selección de variables y procesamiento de espectros, estrategias de fusión de datos y métodos de transferencia (estandarización). / This thesis is focused at developing multivariate analytical screening methodologies for determining the adulteration of culinary spices with Sudan I, II, III and IV dyes. Such dyes are prohibited to be used as additive in foods according to the European legislation because they are Class 3 carcinogens. The proposed methodologies are based on the use of spectroscopic techniques such as UV-Visible, 1H-NMR and Raman along with multivariate data treatment. The applied chemometric tools include the establishment and application of supervised classification techniques combined with exploratory data analysis, data processing and variable selection techniques to extract the maximum possible information from the spectral data. Otherwise some strategies to improve the classification have been evaluated such as data fusion strategies and multivariate transfer (standardization) methods.
|
97 |
A comparison of the applicability and effectiveness of ANOVA with MANOVA for use in the operational evaluation of command and control systemsBurnette, Thomas Nelson 05 1900 (has links)
No description available.
|
98 |
Multicriteria optimization for design of multivariate control charts for manufacturing processesArreola-Risa, Jesus S. 12 1900 (has links)
No description available.
|
99 |
The array-matrix concept- a new approach to multivariate analysis.Tait, George Rodney. January 1971 (has links)
No description available.
|
100 |
Analytical rotation in canonical analysisWong, Eddie Kim January 1990 (has links)
Thesis (Ph. D.)--University of Hawaii at Manoa, 1990. / Includes bibliographical references (leaves 94-95) / Microfiche. / vii, 95 leaves, bound 29 cm
|
Page generated in 0.0538 seconds