by

Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
  • Apr 04 2015 06:54

    jvanbrug on master

    Fix graph annotations (compare)

  • Apr 04 2015 06:41

    jvanbrug on master

    Fix image links in README (compare)

  • Apr 04 2015 06:36

    jvanbrug on master

    Add graphs from python/visualiz… (compare)

  • Apr 04 2015 06:33

    jvanbrug on master

    Sanitize movie titles by removi… (compare)

  • Apr 04 2015 05:52

    jvanbrug on master

    Add test function to python/fac… (compare)

  • Apr 04 2015 05:47

    jvanbrug on master

    Polish python modules and final… (compare)

  • Apr 04 2015 05:33

    jvanbrug on master

    Consolidate and refactor python… (compare)

  • Apr 02 2015 16:13

    obinnae on master

    Refactored code. Got rid of non… Merge branch 'master' of https:… (compare)

  • Apr 02 2015 15:24

    obinnae on cleanup

    Refactored code. Got rid of non… Merge branch 'master' of https:… (compare)

  • Mar 30 2015 22:24

    jvanbrug on matlab

    (compare)

  • Mar 14 2015 19:12

    jvanbrug on master

    Fix user_movie_matrix.csv filep… (compare)

  • Mar 13 2015 21:58

    jvanbrug on master

    Format report (compare)

  • Mar 13 2015 21:43

    jvanbrug on master

    Refactor svd_projection and add… Merge master into local changes (compare)

  • Mar 13 2015 21:34

    OuYamei on master

    final code (compare)

  • Mar 13 2015 21:30

    obinnae on master

    added a heading and our names t… (compare)

  • Mar 13 2015 21:15

    obinnae on master

    first draft of reports (DOC and… (compare)

  • Mar 13 2015 19:28

    obinnae on master

    Added matlab visualization code… (compare)

  • Mar 13 2015 16:33

    jvanbrug on master

    Refactor svd_projection, genera… (compare)

  • Mar 13 2015 08:06

    jvanbrug on new-svd

    (compare)

  • Mar 13 2015 08:05

    jvanbrug on master

    Reorganize repository (compare)

Yamei Ou
@OuYamei
get some result like this
It is after 150 iterations
Just pushed update result and code, however its comment is bad, I will do later if the final result is good enough
obinnae
@obinnae
What lambda and iterations did you use for this recent run?
Jan Van Bruggen
@jvanbrug
@obinnae what are you working on? and in what language?
Jan Van Bruggen
@jvanbrug
@/all i'm going to reorganize the repository
Jan Van Bruggen
@jvanbrug
@OuYamei can you describe what each csv is?
@OuYamei i'll rename them to be more descriptive
Yamei Ou
@OuYamei
ignore the other test data, see the data of 1000 iterations. u.csv v.csv which are the u,v matrix. recover.csv is np.dot(u,v) to compare with user_movie_matrix.csv
Jan Van Bruggen
@jvanbrug
@OuYamei can you check your code in svd_projection and confirm that you calculate u2 and v2 correctly? the two lines both use av, and only u gets transposed
Yamei Ou
@OuYamei
I follow the step-by-step tutorial, there used av for both, I am not sure why
U get transposed is because I want to make the dimension correct otherwise I can not get 2 features for every user.
Jan Van Bruggen
@jvanbrug
For the tutorial you mean his pdf guide?
Yamei Ou
@OuYamei
yes
obinnae
@obinnae
Jan, what movie sequel did you look at besides star wars?
Actually, can you just send me the star wars and Ammityville plots
Ore just post them in the docs since you already have the labels on them
Also, I just realized that I haven't seen most of the movies there besides the disney ones lol
So I wouldn't know any of the movies that it recommends
Jan Van Bruggen
@jvanbrug
Gotcha, I can just do myself then
Yamei Ou
@OuYamei
Jan, after looking for svd, I still think you are correct. however you need to transpose u first.
Jan Van Bruggen
@jvanbrug
yes, i now understand why you reuse av and transpose u
Yamei Ou
@OuYamei
because our notation is not the same where u^T is actually u here. should I change the code, or you've already do that?
Jan Van Bruggen
@jvanbrug
i already did that
i think i pushed it
no i didn't
pushing now
Yamei Ou
@OuYamei
good!
obinnae
@obinnae
I'm going to push my matlab version in a sec
obinnae
@obinnae
Jan, could you include the plot of movies of some similar genre
Jan Van Bruggen
@jvanbrug
absolutely
i found a bug in my visualization code
so the plots might be better
obinnae
@obinnae
That's always good news lol
Also, 1 more plot of 2 different genre's that are still somewhat similar
obinnae
@obinnae
Just insert them in the docs page on the last page when you finish
Jan Van Bruggen
@jvanbrug
you bet!
Jan Van Bruggen
@jvanbrug
@OuYamei I hope your spring break is fun! :smile: I'd like to tidy/clean up the latent factors repository. Which python files are important? I can take care of the TODOs and make the code look pretty, but can you let me know which files/functions actually generated data for our Kaggle submissions? Thanks!
@obinnae can you tidy up the MATLAB code?
Yamei Ou
@OuYamei
I was out of town during spring break and just come back. For mini-project2, I used MatrixFactorization.py to generate data.
How is your break? I just noticed that I missed your party when I was not here.
Jan Van Bruggen
@jvanbrug
So I can delete the other "MatrixFact...blahblah.py" files?
My break was great! I went rock climbing in Joshua Tree National Park with my dad and girlfriend, and I drank lots of quality beers :smile:
Yamei Ou
@OuYamei
Yes
Jan Van Bruggen
@jvanbrug
So I should delete MatrixFact.py and rename MatrixFactoriation.py to MatrixFactorization.py with a t?
sorry, typo, ignore the above
So I should delete MatrixFact.py and rename MatrixFactorizaion.py to MatrixFactorization.py with a t?
Jan Van Bruggen
@jvanbrug
Hey guys, look at the GitHub repo and let me know what you think!
@OuYamei let me know if I deleted any important files or mis-documented anything
You can both fork it if you want!