Abstract: The aim of these papers is to propose a system that helps novice programmers to overcome the syntax errors by using automatic detection and correction technique. This technique is one of the important challenges facing researchers. The proposed system uses Hopfield neural network technic as an associated memory and the one hot vector as a method to represent the data. Where the number of iterations is adopted to reach of the stable state of the network to detect the syntax error and then correct them. The method of representation helps to increase the efficiency of the Hopfield technic and we will discuss this through two experiments. In the first experiment used the development one hot vector method as the matrix to represent the data. While used development one hot vector method as the one vector in the second experiment. The system is trained by using a case study containing a valid syntax statement. While it is tested by generating a set of errors in the statement for the case study. The case study is consists of several examples (Classes and Method) written in the Java programming language. These (Classes and Method) contains statement about the process of the multiplication and the arithmetic summation of the matrix. In the evaluation and testing process, random errors are created that related to a sequence of the tokens in the For-loop, If and While statements. The case study contains (205) statements that include (90) statements of the (For-loop statements) divided into (45) syntax error in the sequence of tokens and (45) correct statement.

Keywords: Correction syntax error, Detection syntax error, One Hot Vector Method, Java.