Commit Graph

31 Commits

Author SHA1 Message Date
Mahdi Dibaiee
6b9cb1fa3e relu: run notmnist using relu activation and draw the chart
[wip] word2vec: work in progress implementation of word2vec
2016-09-13 09:49:44 +04:30
Mahdi Dibaiee
bcc22465d6 fix(xor): better learning rate 2016-09-10 19:21:52 +04:30
Mahdi Dibaiee
f379f208db fix(crossEntropy): implement crossEntropy' to be used in output layer
fix(softmax'): softmax was not correct
2016-09-10 17:43:45 +04:30
Mahdi Dibaiee
c23fd14771 chore: section 2016-09-10 01:22:09 +04:30
Mahdi Dibaiee
6cb130b15e chore(README): simple neural network 2016-09-10 01:20:08 +04:30
Mahdi Dibaiee
58851611dc chore(README): better readme, chart 2016-09-10 01:15:38 +04:30
Mahdi Dibaiee
b26347e19f feat(notmnist): notmnist example using SGD + learning rate decay 2016-09-10 00:36:15 +04:30
Mahdi Dibaiee
ace0a18653 chore(README): explain how the top 10 method increases accuracy and F measure 2016-08-21 01:21:42 +04:30
Mahdi Dibaiee
7d0ce29ba8 chore(README): don't tell them about accuracy, let them try it themselves 2016-08-21 01:07:01 +04:30
Mahdi Dibaiee
25e44d3a8c chore(data): doc-classifier-data 2016-08-21 01:02:45 +04:30
Mahdi Dibaiee
891f48a2d0 feat(topten): top-ten classification with evenly distrubuted data 2016-08-21 00:59:42 +04:30
Mahdi Dibaiee
b2888417bb fix(cleanText): remove unnecessary spaces
fix(run): use `1 - prior` for alpha, no need for smoothing
feat(cleanText): turn all text to lowercase
2016-08-09 16:04:57 +04:30
Mahdi Dibaiee
eebf5e0222 feat(verbose): print more information using -v or --verbose flags 2016-08-08 12:35:26 +04:30
Mahdi Dibaiee
099c25e166 feat(stopwords): removeWords and removeStopwords functions as pre-processors
feat(confidence, WIP): calculate confidence of each classification
2016-08-08 10:02:26 +04:30
Mahdi Dibaiee
ea1f05f001 fix(naivebayes): fix the algorithm to make it actually work
feat(cleanDocuments): preprocess documents, use stemming and stopword elimination for better accuracy
2016-08-05 23:54:36 +04:30
Mahdi Dibaiee
3cf0625794 fix(precision): little bug in implementation 2016-07-30 16:52:34 +04:30
Mahdi Dibaiee
76e7e7faef fix(recall, precision): little bug in calculations
feat(fmeasure): calculate fmeasure using recall and precision
2016-07-29 22:09:30 +04:30
Mahdi Dibaiee
812717522e refactor(&&&): use &&& instead of the arrow function 2016-07-29 18:20:35 +04:30
Mahdi Dibaiee
b5b4629318 feat(results): accuracy, recall and precision functions used to calculate measures
fix: read data from another repository
2016-07-29 17:55:59 +04:30
Mahdi Dibaiee
eeabe4696c chore(README): simple initial README 2016-07-29 16:26:50 +04:30
Mahdi Dibaiee
26eb4531fa feat(naivebayes): implement NaiveBayes algorithm
feat(example): a document classifier using NaiveBayes over reuters data
2016-07-29 16:16:44 +04:30
Mahdi Dibaiee
493a20eb0a feat(crossEntropy): crossEntropy cost function 2016-07-24 10:48:04 +04:30
Mahdi Dibaiee
49606406d1 fix(run): I had written logistic' wrong, that's what happens when you write code while sleepy 2016-07-20 12:06:19 +04:30
Mahdi Dibaiee
b941580273 chore 2016-07-20 10:16:13 +04:30
Mahdi Dibaiee
f5a0c23d99 feat(randomNetwork): generate random networks with a simple function 2016-07-18 17:07:12 +04:30
Mahdi Dibaiee
23851a85f5 fix(stack): use stack build and exec instead of manual stack ghc
refactor: rename from Lib to Sibe
2016-07-18 16:33:34 +04:30
Mahdi Dibaiee
4397f5203a fix(train): fix an error in computing layer's error
feat(examples): add an example (xor)
2016-07-18 00:30:17 +04:30
Mahdi Dibaiee
763faef434 fix(session): missing n definition throws error 2016-07-17 17:14:50 +04:30
Mahdi Dibaiee
fb01d936c2 Merge branch 'master' of github.com:mdibaiee/sibe 2016-07-17 16:54:16 +04:30
Mahdi Dibaiee
59ab00e9c2 initial commit, still work in progress 2016-07-17 16:53:13 +04:30
Mahdi Dibaiee
ba301efb29 Initial commit 2016-07-17 16:52:16 +04:30