This paper intends to review the reasons for the retraction over the last decade. The paper particularly aims at reviewing these reasons with reference to computer science field to assist authors in comprehending the style of writing. To do that, a total of thirty-six retracted papers found on the Web of Science within Jan 2007 through July 2017 are explored. Given the retraction notices which are based on ten common reasons, this paper classifies the two main categories, namely random and nonrandom retraction. Retraction due to the duplication of publications scored the highest proportion of all other reasons reviewed. I. INTRODUCTION Over the years, Steen [1]; Wager and Williams [2]; Ferric C Fang [3]; M. L. Grieneisen and M. Zhang [4]; M. Zhang and M. L. Grieneisen [5]; Daniele Fanelli [6]; E Carafoli [7] Andrew B. Rosenkrantz [8]; and Philippe Mongeon [9] have highlighted that retraction has gradually increased [1]-[9]. Retraction is the act of taking out a paper from scientific publication. The retracted paper defined by the Web of Science (WoS) is "An article that has been withdrawn by an author, institution, editor or a publisher because of an error of unsubstantiated data" 2. Ferric C Fang [3] reasons the Retractions could appear, due to misconduct, fraud, a scientific error, plagiarism, duplicate publication, and so on [3]. It could also be because of faked peer reviews [10]. When this happens, editors will send a notice of concern in a bid for initiating a case of investigation in this respect [11]. Although there is a large proportion of retraction review available in the literature, there has not been much work focusing on the reasons for the act of retraction in computer science. In this paper, an attempt has been done to analyse all computer science retracted papers that were indexed in the WoS between Jan 2007 and July 2017. The pertinent works have taken diverse perspectives on different disciplines. During the last decade, up to 31 th of July 2017, the total number of retracted papers reached 36 in seven areas of computer science. Tentative results have shown that at this current stage, the highest number recorded and contributed to computer science disciplinary accounts for a percentage of 1 This is a pre-print of an in-press article to be published in Advances in Intelligent Systems and Computing by Springer Nature