چكيده به لاتين
Abstract
Natural language processing is one of the important sub-branches of artificial intelligence. Natural language processing is using of computers to process spoken language and written language, that means we can enable computers to analyze and understand speech or written text in form and structure of natural language or produce it.
The sentiment analysis in the text can help determine the authors' emotional comments and goals. Work done on sentiment analysis in Persian language typically focused on identifying positive or negative emotions.
These positive and negative feelings can be divided into six basic feelings, fear, surprise, anger, joy, sadness, disgust. In Persian language, the analysis of these six senses in the texts has not been studied.
In this research, the purpose is to extract the intensity of the fear of Twitter text. In this treatise, it has been tried to pinpoint the severity of fear felt in every Twitter text, for this purpose the texts were measured in different ways, such as the intensity of fear, taking into account the POS(noun, verb, adjective, and adjective)by using Persian vocabulary collection with respect to the role of words and for hashtags in the text of individuals, they were measured using the whole hashtags.
In this regard, 3 Persian vocabulary collections were made and was expanded using the FarsNet Persian vocabulary collection. These three sets of vocabulary using the correctness and review done by the psychologist and the adaptation of this collection of vocabularies was approved by literary texts and was evaluated using supervised learning methods.
In addition to feeling afraid , on the feeling of anger, three vocabulary collections were made in Persian and evaluations were performed on these three sets which made significant results. The final result of the evaluations show that different sets offer different sensitivities for each sensation.
