issues
search
ClimbsRocks
/
machineJS
[UNMAINTAINED] Automated machine learning- just give it a data file! Check out the production-ready version of this project at ClimbsRocks/auto_ml
https://github.com/ClimbsRocks/auto_ml
408
stars
64
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
add in polynomial interactions between features
#129
ClimbsRocks
opened
8 years ago
1
try adding warm_start as a param in makeClassifiers.py
#128
ClimbsRocks
closed
8 years ago
1
Explain to users how to install all python dependencies
#127
ClimbsRocks
closed
8 years ago
1
suppress warnings about logistic regression failing to converge
#126
ClimbsRocks
closed
8 years ago
1
check to make sure validation data is being written correctly
#125
ClimbsRocks
closed
8 years ago
3
bug: not writing trained classifiers to file correctly
#124
ClimbsRocks
closed
8 years ago
1
have a printout of the best stage 0 predictions
#123
ClimbsRocks
opened
8 years ago
0
focus 11-28-15
#122
ClimbsRocks
closed
8 years ago
2
consider bumping down numIterationsPerRound and increasing numRounds commensurately
#121
ClimbsRocks
opened
8 years ago
0
consider warm starting classifier training for the longTraining
#120
ClimbsRocks
closed
8 years ago
1
train up a bunch of weak classifiers
#119
ClimbsRocks
closed
8 years ago
2
give the "un-log" of predicted values for regression problems if relevant
#118
ClimbsRocks
opened
8 years ago
0
write all validationRound predictions to a different folder within predictions
#117
ClimbsRocks
closed
8 years ago
1
make machineJS work by being required
#116
ClimbsRocks
closed
8 years ago
1
support validationRound
#115
ClimbsRocks
closed
8 years ago
1
refactor classifierList
#114
ClimbsRocks
closed
8 years ago
1
fix bug with validation split
#113
ClimbsRocks
closed
8 years ago
1
FUTURE: train algos only on the difficult cases
#112
ClimbsRocks
opened
8 years ago
0
write ALL algos to file, even the crappy ones
#111
ClimbsRocks
closed
8 years ago
1
eventually, take in a --production flag
#110
ClimbsRocks
opened
8 years ago
0
only do long training on an algo if it is within a couple percent of our best
#109
ClimbsRocks
closed
8 years ago
1
only train an algo if it has done alright the first few times
#108
ClimbsRocks
closed
8 years ago
1
clean up code!
#107
ClimbsRocks
closed
8 years ago
1
create bootstrap false algo
#106
ClimbsRocks
closed
8 years ago
1
check out using this wrapper library
#105
ClimbsRocks
opened
8 years ago
0
try new splits of algos based on base_estimator
#104
ClimbsRocks
opened
8 years ago
0
fix issue with --alreadyFormatted
#103
ClimbsRocks
closed
8 years ago
3
create new splits of clRfGini
#102
ClimbsRocks
closed
8 years ago
1
create new splits of AdaBoost
#101
ClimbsRocks
closed
8 years ago
1
update test suite for new 10 rounds default
#100
ClimbsRocks
opened
8 years ago
0
return train_test_split to training.py
#99
ClimbsRocks
closed
8 years ago
1
add tests for label predictions, once i move back to working with one of those datasets again
#98
ClimbsRocks
opened
8 years ago
0
have tinyTest and robustTest as testing options
#97
ClimbsRocks
opened
8 years ago
0
make better tinyTrain.csv dataset for rossman
#96
ClimbsRocks
closed
8 years ago
1
write bestClassifiers to sub-folder named after training file
#95
ClimbsRocks
closed
8 years ago
1
focus
#94
ClimbsRocks
closed
8 years ago
1
move data files out of mjs and into data-for-tests
#93
ClimbsRocks
closed
8 years ago
1
figure out how we're going to get information back out of the running mjs process for testing
#92
ClimbsRocks
closed
8 years ago
1
start testing process by invoking machineJS
#91
ClimbsRocks
closed
8 years ago
1
splitData.py should look for dfValidationSplits.csv
#90
ClimbsRocks
closed
8 years ago
1
take in trainingScore, and then pass that to makePredictions so we can write it to a file and give it to ensembler
#89
ClimbsRocks
closed
8 years ago
1
concat the test predictions and the validation predictions together for each dataset
#88
ClimbsRocks
closed
8 years ago
1
make sure we put the full validation dataset into the new validation folder inside predictions
#87
ClimbsRocks
closed
8 years ago
1
pass in proper folder to ensembler
#86
ClimbsRocks
closed
8 years ago
1
clean: set all default values in a different file than machineJS.js
#85
ClimbsRocks
closed
8 years ago
1
create folders for predictions for each dataset
#84
ClimbsRocks
closed
8 years ago
1
create folders for bestClassifiers for each dataset
#83
ClimbsRocks
closed
8 years ago
1
rename everything in the codebase to use machineJS or mjs
#82
ClimbsRocks
closed
8 years ago
1
idea: randomly shuffling the input data
#81
ClimbsRocks
opened
8 years ago
1
claim npm names
#80
ClimbsRocks
closed
8 years ago
1
Previous
Next