Audience targeting by B-to-B advertisement classification: a neural network approach
From Brede Wiki
|Audience targeting by B-to-B advertisement classification: a neural network approach|
|Authors:||Alan Samuel Abrahams, Eloise Coupey, Eva X. Zhong, Reza Barkhi, Pete S. Manasantivongs|
|Citation:||Expert Systems with Applications 40 (8): 2777-2791. 2013 June|
|Database(s):||Google Scholar cites|
|Web:||Bing Google Yahoo! — Google PDF|
|Article:||BASE Google Scholar PubMed|
|Restricted:||DTU Digital Library|
|Extract:||Talairach coordinates from linked PDF: CSV-formated wiki-formated|
Audience targeting by B-to-B advertisement classification: a neural network approach describes automated text classification in advertissement classification.
- Text sentiment analysis ("sentiment content modeling") with a combination of the ANEW and AFINN word lists.
- General Inquirer was used to label for "semantic conent modeling".
- An artificial neural network is used for classification
The data was captured by 44 undergraduate business major students and contained:
- Advertisement data from Business-to-business magazines
- 5,288 advertisements from July 2007-June 2011
The Data was divided into 7 classes: Capital access, compliance, computer technology, customer service, marketing & innovation, "luxury, travel & personal" and workforce.
Among their results was that a high valance was associated with it being more probably that the advertisement "Marketing & Innovation" group compared to the computer technology group. They believe that it is due to that the Marketing & Innovation group use more possitive words.
- The AFINN word list is sometimes misspelt as AFFIN and the reference is from 2011, not 2012 as stated in the text.
- "The AFFIN dictionary is a major revision of the ANEW dictionary". This is not quite right. The AFINN word list is an independent development and not suppose to be a "revision" of ANEW.