Ecosyste.ms: Issues

An open API service for providing issue and pull request metadata for open source projects.

GitHub / ClimbsRocks/machineJS issues and pull requests

#179 - List index out of range

Issue - State: open - Opened by ittailup over 8 years ago

#178 - Windows multithreading doesn't work

Issue - State: open - Opened by TalhaAsmal over 8 years ago

#177 - Open Validation Indices file for pickling as binary

Pull Request - State: closed - Opened by TalhaAsmal over 8 years ago - 1 comment

#176 - MultinomialNB fails when features have negative values

Issue - State: open - Opened by sunnmy over 8 years ago - 2 comments
Labels: quick win

#175 - Specify full path. Otherwise, machineJS fails when running with preprocessed data.

Pull Request - State: closed - Opened by sunnmy over 8 years ago - 1 comment

#173 - Store validation score in file for ensemble learning

Issue - State: open - Opened by kprimice over 8 years ago - 2 comments

#172 - stop training after a single round if a model is performing very poorly

Issue - State: open - Opened by ClimbsRocks over 8 years ago
Labels: quick win

#171 - Error in installation guide

Issue - State: open - Opened by mmajewsk over 8 years ago - 2 comments

#170 - for the validationRound, write the score from the hyperparameter search, not the final training run

Issue - State: closed - Opened by ClimbsRocks over 8 years ago - 2 comments
Labels: quick win

#169 - investigate how we handle skipping an algo

Issue - State: closed - Opened by ClimbsRocks over 8 years ago - 1 comment
Labels: quick win

#168 - bump down the boost for numRounds for linear models

Issue - State: closed - Opened by ClimbsRocks over 8 years ago - 1 comment
Labels: quick win

#167 - send an email out with updates

Issue - State: open - Opened by ClimbsRocks over 8 years ago

#165 - generally simplify documentation

Issue - State: open - Opened by ClimbsRocks over 8 years ago
Labels: quick win

#164 - update documentation to make it stupid easy to get started with

Issue - State: open - Opened by ClimbsRocks over 8 years ago
Labels: quick win

#163 - add in a table showing an example dataDescription for the .csv files

Issue - State: open - Opened by ClimbsRocks over 8 years ago
Labels: quick win

#162 - tag some of these issues as quick wins and steadily knock them out

Issue - State: closed - Opened by ClimbsRocks over 8 years ago
Labels: quick win

#161 - new installation script

Issue - State: closed - Opened by ClimbsRocks over 8 years ago - 1 comment
Labels: quick win

#160 - Blog Post Series: How to customize machineJS

Issue - State: open - Opened by ClimbsRocks over 8 years ago

#159 - reduce number of times we train each model

Issue - State: closed - Opened by ClimbsRocks over 8 years ago - 1 comment
Labels: quick win

#158 - add in more linear models

Issue - State: open - Opened by ClimbsRocks over 8 years ago
Labels: quick win

#157 - set default to only use linear models for nlp

Issue - State: open - Opened by ClimbsRocks over 8 years ago - 1 comment
Labels: quick win

#156 - shut everything down if there's an error with data-formatter

Issue - State: closed - Opened by ClimbsRocks over 8 years ago - 1 comment
Labels: quick win

#155 - Classifier: output proba instead of category

Issue - State: open - Opened by kprimice over 8 years ago - 1 comment

#154 - Issue after data-formatter

Issue - State: closed - Opened by kprimice over 8 years ago - 2 comments

#153 - Something wrong with data formatter?

Issue - State: open - Opened by TasmaniaKrama over 8 years ago - 11 comments

#152 - Persist everything to disk so a trained instance can be loaded

Issue - State: open - Opened by ClimbsRocks over 8 years ago - 1 comment

#151 - future ideas:

Issue - State: open - Opened by ClimbsRocks over 8 years ago

#150 - allow custom validation splits, like numer.ai uses

Issue - State: closed - Opened by ClimbsRocks over 8 years ago - 1 comment

#149 - Cannot find pySetup/utils.py

Issue - State: open - Opened by eanie over 8 years ago - 20 comments

#148 - TImeseries data?

Issue - State: closed - Opened by JohnnySheffield over 8 years ago - 2 comments

#147 - link in "explaining machineJS to my parents" series

Issue - State: open - Opened by ClimbsRocks almost 9 years ago

#146 - for debugging purposes, load in the raw csv

Issue - State: open - Opened by ClimbsRocks almost 9 years ago

#145 - improve accuracy

Issue - State: open - Opened by ClimbsRocks almost 9 years ago - 1 comment

#144 - have intentional pacing of SIGINT signals

Issue - State: open - Opened by ClimbsRocks almost 9 years ago

#143 - verify that we properly stop training algorithms when they haven't been performing for three rounds

Issue - State: closed - Opened by ClimbsRocks almost 9 years ago - 2 comments
Labels: quick win

#142 - give the user the option of returning a multi-category matrix, not just a single column

Issue - State: closed - Opened by ClimbsRocks almost 9 years ago - 2 comments
Labels: minor focus

#140 - add support for multiple categories in the output column

Issue - State: closed - Opened by ClimbsRocks almost 9 years ago - 2 comments
Labels: major focus

#139 - tweak search parameters

Issue - State: open - Opened by ClimbsRocks almost 9 years ago
Labels: minor focus

#138 - have amazing documentation

Issue - State: open - Opened by ClimbsRocks almost 9 years ago
Labels: major focus

#137 - update verbiage from classifiers to estimators now that we support regression

Issue - State: open - Opened by ClimbsRocks almost 9 years ago
Labels: quick win

#136 - packages to install in installation script

Issue - State: closed - Opened by ClimbsRocks almost 9 years ago - 1 comment

#135 - create installation script

Issue - State: closed - Opened by ClimbsRocks almost 9 years ago - 1 comment

#134 - write all trained algo's parameters to file

Issue - State: open - Opened by ClimbsRocks almost 9 years ago
Labels: minor focus, quick win

#133 - only run the MLP if the user has ^v.0.18.0 of sklearn installed

Issue - State: closed - Opened by ClimbsRocks almost 9 years ago - 1 comment

#129 - add in polynomial interactions between features

Issue - State: open - Opened by ClimbsRocks almost 9 years ago - 1 comment

#128 - try adding warm_start as a param in makeClassifiers.py

Issue - State: closed - Opened by ClimbsRocks almost 9 years ago - 1 comment

#127 - Explain to users how to install all python dependencies

Issue - State: closed - Opened by ClimbsRocks almost 9 years ago - 1 comment

#126 - suppress warnings about logistic regression failing to converge

Issue - State: closed - Opened by ClimbsRocks almost 9 years ago - 1 comment

#125 - check to make sure validation data is being written correctly

Issue - State: closed - Opened by ClimbsRocks almost 9 years ago - 3 comments

#124 - bug: not writing trained classifiers to file correctly

Issue - State: closed - Opened by ClimbsRocks almost 9 years ago - 1 comment

#123 - have a printout of the best stage 0 predictions

Issue - State: open - Opened by ClimbsRocks almost 9 years ago

#122 - focus 11-28-15

Issue - State: closed - Opened by ClimbsRocks almost 9 years ago - 2 comments

#120 - consider warm starting classifier training for the longTraining

Issue - State: closed - Opened by ClimbsRocks almost 9 years ago - 1 comment

#119 - train up a bunch of weak classifiers

Issue - State: closed - Opened by ClimbsRocks almost 9 years ago - 2 comments

#117 - write all validationRound predictions to a different folder within predictions

Issue - State: closed - Opened by ClimbsRocks almost 9 years ago - 1 comment

#116 - make machineJS work by being required

Issue - State: closed - Opened by ClimbsRocks almost 9 years ago - 1 comment

#115 - support validationRound

Issue - State: closed - Opened by ClimbsRocks almost 9 years ago - 1 comment

#114 - refactor classifierList

Issue - State: closed - Opened by ClimbsRocks almost 9 years ago - 1 comment

#113 - fix bug with validation split

Issue - State: closed - Opened by ClimbsRocks almost 9 years ago - 1 comment

#112 - FUTURE: train algos only on the difficult cases

Issue - State: open - Opened by ClimbsRocks almost 9 years ago

#111 - write ALL algos to file, even the crappy ones

Issue - State: closed - Opened by ClimbsRocks almost 9 years ago - 1 comment

#110 - eventually, take in a --production flag

Issue - State: open - Opened by ClimbsRocks almost 9 years ago

#109 - only do long training on an algo if it is within a couple percent of our best

Issue - State: closed - Opened by ClimbsRocks almost 9 years ago - 1 comment

#108 - only train an algo if it has done alright the first few times

Issue - State: closed - Opened by ClimbsRocks almost 9 years ago - 1 comment

#107 - clean up code!

Issue - State: closed - Opened by ClimbsRocks about 9 years ago - 1 comment

#106 - create bootstrap false algo

Issue - State: closed - Opened by ClimbsRocks about 9 years ago - 1 comment

#105 - check out using this wrapper library

Issue - State: open - Opened by ClimbsRocks about 9 years ago

#104 - try new splits of algos based on base_estimator

Issue - State: open - Opened by ClimbsRocks about 9 years ago

#103 - fix issue with --alreadyFormatted

Issue - State: closed - Opened by ClimbsRocks about 9 years ago - 3 comments

#101 - create new splits of AdaBoost

Issue - State: closed - Opened by ClimbsRocks about 9 years ago - 1 comment

#99 - return train_test_split to training.py

Issue - State: closed - Opened by ClimbsRocks about 9 years ago - 1 comment

#94 - focus

Issue - State: closed - Opened by ClimbsRocks about 9 years ago - 1 comment

#82 - rename everything in the codebase to use machineJS or mjs

Issue - State: closed - Opened by ClimbsRocks about 9 years ago - 1 comment

#76 - explore using dense arrays as an option

Issue - State: closed - Opened by ClimbsRocks about 9 years ago - 1 comment

#72 - switch gridsearchcv to be randomsearch

Issue - State: closed - Opened by ClimbsRocks about 9 years ago - 2 comments

#70 - better parameters for xgboost for gridsearch

Issue - State: open - Opened by ClimbsRocks about 9 years ago - 1 comment
Labels: minor focus

#69 - new classifiers: Bagging Estimator

Issue - State: open - Opened by ClimbsRocks about 9 years ago
Labels: minor focus

#68 - new classifiers: lda and qda

Issue - State: open - Opened by ClimbsRocks about 9 years ago
Labels: minor focus

#67 - new classifiers: naive bayes

Issue - State: closed - Opened by ClimbsRocks about 9 years ago - 1 comment

#64 - nns: have the nn itself figure out when to stop training

Issue - State: closed - Opened by ClimbsRocks about 9 years ago - 1 comment

#62 - new classifiers: many more splits of the nn

Issue - State: closed - Opened by ClimbsRocks about 9 years ago - 1 comment

#56 - generalize all paths to use path module, rather than forward or back slashes

Issue - State: open - Opened by ClimbsRocks about 9 years ago
Labels: minor focus

#53 - installation steps/requirements/dependencies

Issue - State: closed - Opened by ClimbsRocks about 9 years ago - 4 comments

#43 - add .gitkeep to all folders we are going to need

Issue - State: closed - Opened by ClimbsRocks about 9 years ago

#42 - whenever possible, split classifiers, don't combine

Issue - State: closed - Opened by ClimbsRocks about 9 years ago - 1 comment

#40 - fix up the neural network; right now it's not doing too well

Issue - State: closed - Opened by ClimbsRocks about 9 years ago - 1 comment

#39 - documentation- getting started

Issue - State: closed - Opened by ClimbsRocks about 9 years ago - 1 comment

#36 - training, testing, training

Issue - State: closed - Opened by ClimbsRocks about 9 years ago - 1 comment

#28 - write everything to a file

Issue - State: open - Opened by ClimbsRocks about 9 years ago
Labels: major focus

#24 - update documentation!

Issue - State: closed - Opened by ClimbsRocks about 9 years ago - 3 comments

#15 - add support for multiple output columns

Issue - State: open - Opened by ClimbsRocks about 9 years ago

#13 - figure out how to make this more generally useful for passing in non-Kaggle data

Issue - State: open - Opened by ClimbsRocks over 9 years ago
Labels: major focus