Exploratory Correlation Analysis

نویسندگان

  • Jos Koetsier
  • Donald MacDonald
  • Darryl Charles
  • Colin Fyfe
چکیده

We present a novel unsupervised artificial neural network for the extraction of common features in multiple data sources. This algorithm, which we name Exploratory Correlation Analysis (ECA), is a multi-stream extension of a neural implementation of Exploratory Projection Pursuit (EPP) and has a close relationship with Canonical Correlation Analysis (CCA). Whereas EPP identifies ”interesting” statistical directions in a single stream of data, ECA develops a joint coding of the common underlying statistical features across a number of data streams. It has been shown that the principle of contextual guidance may be used to find a sparse coding of the features in dual natural image patches that is very different from single stream sparse coding experiments. The network only identifies those features which exist in both data streams and thus tend to be fewer in number and more complex in nature.

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

Subjective Evaluation of Mobile Speech Input: Experiences from Two Case Studies

Thorough understanding of subjective and objective measurements of speech-based interaction, especially of its user experience, is vital for practical application development. We present findings from two case studies where multimodal applications containing speech input were evaluated using a subjective evaluation methodology. Responses were investigated using both correlation analysis and exp...

متن کامل

Developing a Critical Checklist for Textbook Evaluation

This study has been carried out to develop a critical checklist for global/commercial textbooks which play a crucial role in language teaching/learning. For this aim, a number of items have been developed based on a comprehensive review of the related literature and experts’ opinions. The tentative checklist was administered to the targeted population, yet 326 checklists were deemed appropriate...

متن کامل

Immediate Reward Reinforcement Learning for Projective Kernel Methods

We extend a reinforcement learning algorithm which has previously been shown to cluster data. We have previously applied the method to unsupervised projection methods, principal component analysis, exploratory projection pursuit and canonical correlation analysis. We now show how the same methods can be used in feature spaces to perform kernel principal component analysis and kernel canonical c...

متن کامل

Canonical correlation analysis of online video advertising viewing motivations and access characteristics

This study investigates consumers’ motivations for watching online video ads and the relationship between the motivations and access characteristics of viewers. Uses and gratifications and characteristics of consumers’ exposure to online video ads were reviewed for this study. First, findings of exploratory factor analysis revealed five different motivations for viewing online video ads – socia...

متن کامل

Canonical Correlation & Principal Components Analysis

Canonical Correlation is one of the most general of the multivariate techniques. It is used to investigate the overall correlation between two sets of variables (p’ and q’). The basic principle behind canonical correlation is determining how much variance in one set of variables is accounted for by the other set along one or more axes. If there is more than one axis, they must be orthogonal. Un...

متن کامل

University of Pittsburgh Cs 2750 Machine Learning Handout 3 Professor Milos Hauskrecht Solutions to Problem Set 3 Problem 1. Linear Regression Part 1. Exploratory Data Analysis

(a) Attribute 4, CHAS, is the only binary attribute. (b) Attribute 13—LSTAT—has the highest negative correlation and attribute 6—RM—has the highest positive correlation with the target attribute.

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

عنوان ژورنال:

دوره   شماره 

صفحات  -

تاریخ انتشار 2002