Code completion with statistical language models

Key Points

• The aim is to synthesize effective code completions for programming tasks involving APIs.
• Developed statistical language models for predicting method call sequences.
• Tested model performance on various programming scenarios.
• Evaluated the accuracy of code completions generated for holes in programs.
• Achieved high accuracy rates in predicting method calls with statistical models.
• Demonstrated significant improvement in code completion quality over traditional methods.