International Journal of Scientific & Engineering Research

-Programmers frequently write code that can be similar to existing code which may be written somewhere. An approach could help programmers to complete partially written code snippets to implement necessary functionality, help to discover extended lines of code to the partial code which are commonly used statements by other programmers, help to cross-check against similar code written by other programmers and also helps in fixing common mistakes and errors. Although there are many existing techniques which could potentially be used to get code suggestions. For example, code-to-code search tools could retrieve relevant code snippets from a corpus using a partial code snippet as query. However, the challenging task is to locate relevant source code from the large size code repositories and returns lots of relevant code snippets without removing or aggregating similar-looking ones. So, a new approach for code suggestion can be developed which takes a code snippet (partial code) as input and which can recommend several succinct source codes. The tool performs manual filtering from the dataset available and predicts the top five similar codes based on the similarity score.