To provide a rich experience to Octave users, the command line suggestion feature, aims to be a fast and accurate method to suggest corrections to the users whenever they make a typographic error while working on the Octave command window. It should also be able to provide the users, an easy method to select and execute the corrections suggested and enable/disable the feature. This could get tricky as Octave comes in both a GUI and a CLI. Therefore, for a successful integration of this feature with Octave, understanding how its parser, error handling, and internal functions work would be essential. The next part of the problem would be to generate the words to suggest. A trivial solution of this problem is easy but computationally expensive and hence not feasible. The challenging part therefore, is to minimize the speed-accuracy trade-off of the feature. A starting point for that would be a smart (probability based for example) implementation of the Edit-distance algorithm, using dynamic programming. Later, a Neural Network based idea would also be tested. The goal of-course, is to use the algorithm that best minimizes the speed-accuracy trade-off and provides the best UX.

Organization

Student

P SUDEEPAM

Mentors

  • Douglas Stewart
  • Nicholas Jankowski
close

2018