## page was renamed from 機械学習 #pragma section-numbers off [[TableOfContents]] = 概要 = 機械学習 = インストール = http://conda.pydata.org/miniconda.html = 関連ライブラリ = * ["NumPy"] * ["Matplotlib"] * ["scikit-learn"] * ["pandas"] * ["word2vec"] [https://github.com/OryxProject/oryx OryxProject/oryx · GitHub] = ディープラーニング = * ["DeepLearning"] * ["TensorFlow"] * ["Chainer"] * ["Darknet"] = 教材 = * [https://github.com/jakevdp/2013_fall_ASTR599/ jakevdp/2013_fall_ASTR599 · GitHub] * [https://github.com/jakevdp/sklearn_pycon2015 jakevdp/sklearn_pycon2015 · GitHub] = 形態素解析 = * [http://www.mwsoft.jp/programming/munou/mecab_nitteretou.html 日本テレビ東京で学ぶMeCabのコスト計算 | mwSoft] * [http://ukonlly.hatenablog.jp/entry/20100831/1283246213 wikipedia dump を使って複合名詞を判定してみる - 開発めも２] * [http://diary.overlasting.net/2015-03-13-1.html MeCab 用の新語辞書 mecab-ipadic-neologd を公開しました] = 文章解析 = * [https://github.com/anttttti/Wordbatch GitHub - anttttti/Wordbatch: Parallel text feature extraction for machine learning] = アルゴリズム等 = == 辞書 == * [http://hivecolor.com/id/65 tfidf、LSI、LDAの違いについて調べてみた - Hive Color] * [http://hivecolor.com/id/88 tfidf, lsi, ldaを使ったツイッターユーザーの類似度計算 - Hive Color] == SVM == * [http://qiita.com/pika_shi/items/5e59bcf69e85fdd9edb2 SVMを使いこなす！チェックポイント8つ] * [http://qiita.com/sz_dr/items/f3d6630137b184156a67 SVM(RBFカーネル)のハイパーパラメータを変えると何が起こるの?] == 不均衡データ == * [http://www.slideshare.net/sfchaos/ss-11307051 不均衡データのクラス分類] * [http://d.hatena.ne.jp/dichika/20130424/p1 Crowdsolvingに参加して7/42位だった] * SMOTE * Granular Support Vector Machines == multiclass-multilabel == * [http://stackoverflow.com/questions/10526579/use-scikit-learn-to-classify-into-multiple-categories python - use scikit-learn to classify into multiple categories - Stack Overflow] == Active Learning == * [http://www.slideshare.net/shuyo/introduction-to-active-learning-25787487 Active Learning 入門] == 未分類 == * [http://kazoo04.hatenablog.com/entry/2012/12/20/000000 オンライン線形分類器とSCW - Sideswipe] = オープンデータ = * [https://www.csie.ntu.edu.tw/~cjlin/libsvmtools/datasets/ LIBSVM Data: Classification, Regression, and Multi-label] * [https://qiita.com/tmp_llc/items/7296c5d6bb8769b18d24 オープンデータ取得先まとめ - Qiita] = 表示 = pd.crosstab * [http://stackoverflow.com/questions/5821125/how-to-plot-confusion-matrix-with-string-axis-rather-than-integer-in-python matplotlib - How to plot confusion matrix with string axis rather than integer in python - Stack Overflow]:confusion matrix のグラフを綺麗な画像にする * https://github.com/wellflat/cat-fancier/blob/master/classifier/bin/report.py * https://github.com/alxndrkalinin/mi-pred/blob/master/ML.py * http://akiniwa.hatenablog.jp/entry/2013/08/11/182756 * http://qiita.com/wellflat/items/0b6b859bb275fd4526ed = グラフ = * http://bokeh.pydata.org/en/latest/index.html * https://github.com/mwaskom/seaborn * [http://matplotlib.org/examples/color/colormaps_reference.html color example code: colormaps_reference.py — Matplotlib 1.4.2 documentation] * http://nbviewer.ipython.org/github/bugra/pydata-nyc-2014/blob/master/3.%20Scikit%20Learn%20-%20Supervised%20Learning.ipynb * http://bokeh.pydata.org/en/latest/docs/gallery/cat_heatmap_chart.html * http://isaacslavitt.com/2014/10/24/spdc-lightning-talk/ * http://nbviewer.ipython.org/github/jakevdp/ESAC-stats-2014/blob/master/notebooks/05.2-Clustering-KMeans.ipynb * http://www.interaction-ipsj.org/archives/paper2013/data/Interaction2013/oral/data/pdf/13INT001.pdf = AUC = * http://scikit-learn.org/stable/auto_examples/plot_roc.html * http://scikit-learn.org/stable/auto_examples/plot_roc_crossval.html * https://gist.github.com/coreylynch/4150976 * http://www.hpl.hp.com/techreports/2003/HPL-2003-4.pdf * http://otndnld.oracle.co.jp/document/products/oracle10g/102/doc_cd/datamine.102/B19263-01/3predictive.htm = Amazon AWS ノウハウ = * [https://speakerdeck.com/satorukadowaki/aws-apigateway-plus-python-lambda-plus-neologddezuo-rusabaresuri-ben-yu-xing-tai-su-jie-xi-api AWS APIGateway + Python Lambda + NEologdで作るサーバレス日本語形態素解析API // Speaker Deck] * [https://speakerdeck.com/kumon/maikurosabisuapurikesiyontositefalseji-jie-xue-xi マイクロサービスアプリケーションとしての機械学習 // Speaker Deck] = R関連 = * [https://atnd.org/events/22039 R Advent Calendar 2011] * [https://atnd.org/events/31973 R Advent Calendar 2012] * [https://atnd.org/events/45043 R Advent Calendar 2013] * [https://atnd.org/events/58648 R Advent Calendar 2014] = 用語メモ = カッパ係数 kappa、自由回答，インタビューなどの分類の信頼性，一致度 http://www.mizumot.com/stats/kappa.htm = その他資料 = * [https://speakerdeck.com/mathetake/yusaxing-dong-falseshu-li-moteruto-gao-su-tui-jian-sisutemu ユーザー行動の数理モデルと 高速推薦システム // Speaker Deck] * [https://speakerdeck.com/hoxom/sgmrfmix 多次元時系列の異常検知手法 sGMRFmix について /sGMRFmix // Speaker Deck] = 未分類 = [http://www.shojiro-tanaka.net/gradOutputs/2004/saitoH.pdf 機械学習の評価] [http://universityofbigdata.net/competition/tutorial/5681717746597888 テキスト分類問題その１ チュートリアル] [https://github.com/notani/uob/tree/master/03_text uob/03_text at master · notani/uob · GitHub] [https://www.udemy.com/computervision/ ベイズ推定とグラフィカルモデル：コンピュータビジョン基礎1 - Udemy] [https://github.com/CamDavidsonPilon/Probabilistic-Programming-and-Bayesian-Methods-for-Hackers CamDavidsonPilon/Probabilistic-Programming-and-Bayesian-Methods-for-Hackers · GitHub] = 参考サイト = ---- CategoryTechnical