Dimensional Geometry , Curse of Dimensionality , Dimension Reduction
نویسنده
چکیده
High-dimensional vectors are ubiquitous in algorithms and this lecture seeks to introduce some common properties of these vectors. We encounter the so-called curse of dimensionality which refers to the fact that algorithms are simply harder to design in high dimensions and often have a running time exponential in the dimension. We also show that it is possible to reduce the dimension of a dataset sometimes —and for some purposes. Notation: For a vector x ∈ <n its `2-norm is |x|2 = ( ∑ i x 2 i ) 1/2 and the `1-norm is |x|1 = ∑ i |xi|. For any two vectors x, y their Euclidean distance refers to |x− y|2 and Manhattan distance refers to |x− y|1. We start with some useful generalizations of geometric objects to higher dimensional geometry:
منابع مشابه
Lecture 6 : Curse of Dimensionality , Dimension Reduction
High dimensional vectors appear frequently in recent development in CS, examples of which are user-movie ratings of netflix, DNA strings of patients, and images pixel values. In this lecture we study higher dimensions geometry and see how randomization help us to design algorithm. We first start with a few definitions, then we discuss several properties of high dimensional geometry, and finally...
متن کاملGeometric Optimization April 12 , 2007 Lecture 25 : Johnson Lindenstrauss Lemma
The topic of this lecture is dimensionality reduction. Many problems have been efficiently solved in low dimensions, but very often the solution to low-dimensional spaces are impractical for high dimensional spaces because either space or running time is exponential in dimension. In order to address the curse of dimensionality, one technique is to map a set of points in a high dimensional space...
متن کاملIncomplete Pivoted QR-based Dimensionality Reduction
High-dimensional big data appears in many research fields such as image recognition, biology and collaborative filtering. Often, the exploration of such data by classic algorithms is encountered with difficulties due to ‘curse of dimensionality’ phenomenon. Therefore, dimensionality reduction methods are applied to the data prior to its analysis. Many of these methods are based on principal com...
متن کاملEncoding and Decoding Techniques for Distributed Data Storage Systems
Dimensionality reduction is the conversion of highdimensional data into a meaningful representation of reduced data. Preferably, the reduced representation has a dimensionality that corresponds to the essential dimensionality of the data. The essential dimensionality of data is the minimum number of parameters needed to account for the observed properties of the data [4]. Dimensionality reducti...
متن کاملClutter Reduction in Multi-Dimensional Visualization by Using Dimension Reduction
The volume of Big data is increasing in gigabytes day by day which are hard to make sense and difficult to analyze. The challenges of big data are capturing, storing, searching, sharing, analysis and visualization of these datasets. Big data leads to clutter in their visualization. Clutter is a crowded or disordered collection of graphical entities in information visualization. It can blur the ...
متن کامل