The type of telecommunication call is known by the phone number and the behavior of the call. By examining the rules and patterns of the call, the detection of the fraudulent phone call can be done. Scam calls have become a harmful issue due to which people are losing money in recent years. A simple and efficient way to tackle this problem involves blacklisting fraudulent phone numbers. This method is used by the existing scam call detection apps. These apps use the user reviews and black-list the fraudulent numbers and block them with the permission of the user. However, scammers use software to mask their phone numbers and to display their numbers as official numbers of trustworthy organizations. Such software can be easily morphed by scammers. Thus, the existing methods of scam call detection have proven to be inefficient. In this paper, we have made a comparison between two papers that have a solution for this problem. One uses Natural Language Processing(NLP), in which Machine Learning algorithms are used to examine the data thoroughly and to use the results of the previously collected data to build the classes of data. NLP is used to extract the characteristics of textual data. In the other paper on the Deep Learning(DL) approach, a classifier is built on the basis of records of the call details for fraud call recognition. This DL approach is accurate and effective as it is capable of building similar features of the phone number and does proper recognition of fraudulent calls. Whereas, a simple Natural Language Processing and Machine Learning approach are also equally powerful.