Advertisements are the most powerful means for communicating the marketing message to the target audience. The presence of likeable attributes in ads has profound effect on the mindset of the audience and results in creating a positive image about the ads and consequently, the brands. This article focuses on understanding and using likeability in television commercials.

Extracting useful information from the volume of data being generated at a phenomenal rate is an important and challenging issue at present. The rough set theory, a new mathematical approach to data analysis, based on classification, is a promising technique. In real world situations, collected data is usually uncertain and incomplete. For this reason, this theory alone is not sufficient to properly describe the information hidden in a data set (Magnani, 2003). To deal with complex data, the hybridization of rough set theory with neural networks, can be a useful technique (Magnani). In this paper, a hybrid approach of expert system with a reduct algorithm in order to generate effective reduced data set (reduct) and to construct the core of the attribute set has been proposed. In this approach, the reduced data set is transformed suitably to input neural network to improve classification accuracy. The proposed algorithm is implemented on medical data set to test the efficiency of the algorithm.

The rough set theory, introduced by Pawlak (1982), emerged about 25 years ago and is now a rapidly developing branch of artificial intelligence and soft computing. On the other hand, artificial neural network is considered one of the most powerful predictors for making inferences from partial information. This paper presents some approaches to incorporating rough set and neural network methods into one system for decision or classification support. It is interesting to note that these two methodologies try to incorporate both the approaches in one combined system. Both methodologies have their place among intelligent classification and decision support methods. In this method, the pre-processing part uses rough set theory and the classification part uses neural network. This connectionist model can be especially good at handling noisy and partial information. The paper attempts to preprocess input data for a neural network with rough set methods. Further, related rough set concepts are briefly introduced; the concept of incorporating rough sets methods into construction of neural network by using rough neurons is also described, and the paper concludes with a discussion of the usage of rough set methods and knowledge gained from them in the process of establishing the architecture and initial state of a neural network for a given problem.

Intelligent Classification Using Rough Sets and Neural Networks, network, information, elements, system, attributes, decision, equivalence, methodologies, system, technique, developing, efficiency, function, knowledge, classification, combination, computing, translating, classification, variables, enhancing, performance, PreProcessing