چكيده به لاتين
In recent years, the explosive growth of online media, such as blogs and social networking sites, has enabled costumers to write and share their personal purchasing experiences and opinions with other people. Buying behavior can be influenced by costumer attitude toward products brands and features. Opinion analysis can be used for identifying factors impacting on buying behavior. Traditional Opinion analysis methods cannot be useful because need huge volume of information, variety and quick response. We propose a novel approach for costumer behavior prediction based on attitude dimensions analysis by using text mining techniques and content mining in social network in three popular social media.We investigate automobile industries for six luxury and famous brands and also four electronic product brands as a case study. We addressed costumer buying behavior Results show that attention to costumer attitude can predict fairly accurate future behavior in automobile industries .We use n-grams method for generation feature vectors and machine learning approach for classification as a multi-approach. Results show positive behavior more stems by cognitive behavioral than emotional behavior and costumer follow more cognition opinion than emotional. We also test this approach in electronic products and results show that this model cannot predict costumer behavior well. In this case, Costumer behavior is not dependent on costumer attitude. Therefore, it is concluded that in the automobile industries costumers buying behavior is based on the costumer attitude toward the goods and is not dependent on electronics industries. Many factors can influence on costumer buying behavior like iterative, exciting buying behavior that Due to limitations not point in this research.
Keywords: Costumer behavior, Text mining, social network, Attitude dimension, Behavior prediction
