خط مشی دسترسیدرباره ما
ثبت نامثبت نام
راهنماراهنما
فارسی
ورودورود
صفحه اصلیصفحه اصلی
جستجوی مدارک
تمام متن
منابع دیجیتالی
رکورد قبلیرکورد بعدی
Document Type : Latin Dissertation
Language of Document : English
Record Number : 149252
Doc. No : ET21044
Main Entry : Sajama
Title Proper : Nonparametric methods for learning from data
Note : This document is digital این مدرک بصورت الکترونیکی می باشد
Abstract : Developing statistical machine learning algorithms involves making variousdegrees of assumptions about the nature of the data being modeled. Nonparametricmethods are useful when prior information regarding the parametricform of the model is unavailable or invalid. This thesis presents non-parametricmethods for tackling various modeling requirements.The first part of this thesis presents a pair of unsupervised and supervisedlinear dimensionality reduction techniques that are suitable for various datatypes like binary and integer along with real-valued data. They are based on asemi-parametric mixture of exponential family distributions where no parametricassumptions are made about the latent distribution and the parametric form of thenoise distribution is to be chosen based on the data type, for example Bernoulli forbinary data, etc. A key feature of.
Subject : Electericl tess
: برق
electronic file name : TL44192.pdf
Title and statement of responsibility and : Nonparametric methods for learning from data [Thesis]
 
 
 
(در صورت عدم وضوح تصویر اینجا را کلیک نمایید)