Skip to content
Commits on May 23, 2012
  1. Unfinished work!

    Work on adding small video tutorials.
    Also, make evaluation of hint sequences later.
    committed May 23, 2012
  2. Merge branch 'master' of github.com:bvds/andes

    Conflicts:
    	LogProcessing/CognitiveModels/self-improved-tutor.tex
    committed May 22, 2012
Commits on May 22, 2012
  1. Add list of features.

    committed May 21, 2012
Commits on May 21, 2012
  1. Fix broken IEA-main hint. Make hint sequence

    no longer takes tutor turn structs:  supply function instead.
    committed May 21, 2012
Commits on May 20, 2012
  1. Added list of features used.

    committed May 19, 2012
Commits on May 19, 2012
Commits on May 17, 2012
  1. Finish up improvements to objective function.

    Start writing up result.
    committed May 17, 2012
Commits on May 15, 2012
  1. Don't test for validity when doing objective function,

    make validity test average over all submodels.
    validity test not debugged yet.
    committed May 15, 2012
Commits on May 14, 2012
  1. Add paper temporarily.

    committed May 14, 2012
Commits on May 11, 2012
  1. Contour plots!

    committed May 11, 2012
Commits on May 10, 2012
  1. Use discount factor based on number of incorrect

    steps for this kc.  Work on recursion relation for
    probability of learning integral (unstable).
    committed May 10, 2012
  2. Fix up language for steps and transactions.

    Remove unused code in step-model.php and change code
    to use probability of learning gain.  Still need to test scripts.
    committed May 9, 2012
Commits on May 9, 2012
  1. Added code for calculating probability of learning.

    Not tested or used yet.  Should test against mathematica code.
    committed May 8, 2012
  2. Figured out right way to determine chance that

    learning has occurred.
    committed May 8, 2012
Commits on May 8, 2012
  1. Trying to decide if it is ok to use expectation

    values and variances (for the error) to determine if
    learning has occurred.
    committed May 8, 2012
Commits on May 2, 2012
  1. Fix probabilities to include no-learning fit.

    Use average over models in a consistant manner.
    Include learning gains and discount factor in reward
    function.  Remove term with negative rewards.
    committed May 1, 2012
Commits on Apr 30, 2012
  1. @akondave
Commits on Apr 26, 2012
Commits on Apr 24, 2012
  1. Count all turns in opportunity when waiting the

    importance of a term to learning or for a slip.
    Remove "none" from list of KC's.
    committed Apr 24, 2012
Commits on Apr 23, 2012
Commits on Apr 21, 2012
  1. Take log of some features.

    Mathematica notebook to calculate best model policy
    using perceptron approach.
    committed Apr 20, 2012
Commits on Apr 18, 2012
  1. Print instances of reward, policy used, and pointer to associated state

    in cases where policy may apply.
    committed Apr 17, 2012
Commits on Apr 17, 2012
  1. Print out assignment of blame chronologically

    for each student.  Fix routine returning current flounder time.
    committed Apr 17, 2012
Commits on Apr 16, 2012
  1. Print out student state and add some more observables

    Move student state code to separate file.
    committed Apr 16, 2012
Commits on Apr 14, 2012
  1. Start work on creating student state.

    Mark steps without KC assigned as KC=none.
    Do assignment of blame for steps without object
    (time heuristic) in real time.
    committed Apr 13, 2012
Commits on Apr 12, 2012
  1. Move analysis of training video and fade problems

    to separate file.
    committed Apr 12, 2012
Commits on Apr 8, 2012
Commits on Apr 7, 2012
Commits on Apr 6, 2012
  1. Fix bug for confusedtID.

    committed Apr 6, 2012
Commits on Apr 5, 2012
  1. Use library routine for determining floundering

    and time on task.
    committed Apr 5, 2012
Commits on Mar 30, 2012
  1. Had two error tests with name "default-wrong-dir".

    Add warning for redefining operator.
    committed Mar 30, 2012
Something went wrong with that request. Please try again.