確çã»çµ±è¨
ãã£ãªã¯ã¬(Dirichlet)åå¸ã3Dã§å¯è¦åãã ï¼3次å ã®ï¼ãã£ãªã¯ã¬åå¸ãpythonã®matplotlib.plot_surfaceã§å¯è¦åããã¡ã¢ï¼ ãã£ãªã¯ã¬(Dirichlet)åå¸ã3Dã§å¯è¦åãã ç°å¢ ãã£ãªã¯ã¬åå¸ matplotlibã§å¯è¦åããï¼ åèãªã³ã¯ MathJax.Hub.Config({ tâ¦
[é¸æçµ±è¨]ã©ã®å æ´¾ãå°æ± ç¾ååæ°ãåé¸ããããããã¤ãºæ¨å®ãã é½ç¥äºé¸ã2020å¹´7æ5æ¥ã«ããã¾ããï¼ å¤§å·®ã§å°æ± ç¾ååæ°ãåé¸ãã¾ããï¼ ä»åã¯å°æ± ç¾ååæ°ã«æ票ããå ã©ã®å æ´¾ãå¤ããããã¤ãºæ¨å®ã§è¨ç®ãã¾ãï¼ çµè«ããããã¨ç¡å 派層ï¼å°æ± æ°â¦
å åç¹ã¨éã¿ä»ãå¹³åã®èå¯ å åç¹ã¨éã¿ä»ãå¹³åã®èå¯ã¡ã¢ gist: å åç¹ã¨éã¿ä»ãå¹³å · GitHub å åç¹ã¨éã¿ä»ãå¹³åã®èå¯ å åç¹1ã¤ã®å ´å ãªãéã¿ãéã«ãªãã®ã å åç¹ã2ã¤ããå ´å ç¹°ãè¿ãå åç¹ãåããå ´å éã¿ä»ãå¹³å ãã¼ã¿ã2種é¡ã®å ´åâ¦
ã¬ã³ãåå¸ã®ãã©ã¡ã¼ã¿æ¨å® ã¬ã³ãåå¸ã®ãã©ã¡ã¼ã¿æ¨å®ã¡ã¢ï¼ 以ä¸ã®ææ³ã®ã¡ã¢ æ¨æ¬å¹³åï¼æ¨æ¬åæ£ããã®è¿ä¼¼å¤æ¨å® æå°¤æ¨å® shapeãã©ã¡ã¼ã¿ã®è¿ä¼¼å¤æ¨å® ãããåæå¤ã¨ãã¦æ°å¤è§£æçã«ãã¥ã¼ãã³ã»ã©ãã½ã³æ³ã§æ±ãã MathJax.Hub.Config({ tex2jax:â¦
ãªãã¸å帰(Ridge Regression)ã®æ£è¦æ¹ç¨å¼ãå°åº ãªãã¸å帰(Ridge Regression)ã®æ£è¦æ¹ç¨å¼ãå°åºããã¡ã¢ï¼ L2ãã«ã ã®2ä¹ã«ããå¾®åã容æã«ã§ããã®ã§è§£æ解ã¨ãã¦å°åºã§ããï¼ MathJax.Hub.Config({ tex2jax: { inlineMath: [['$','$'], ['\\(','\\)']â¦
ã«ã¤ã¸ã®Eã«ã¼ãã®åçãè¨ç®ãã å³æ¸é¤¨ã§Newtonå¥åâ確çã«å¼·ããªãâãå¶ç¶ãã«ã²ããæ°å¦æ³åâï¼2010å¹´çºè¡ãªã®ã§ã¡ãã£ã¨å¤ãï¼ãèªãã§ããé¢ç½ãã³ã©ã ããã£ãï¼ é¸æå¶ãããããï¼ã¤ã¾ãEã«ã¼ãã®åçã§ããï¼ ç¢ºçã«å¼·ããªãâãå¶ç¶ãã«ã²ããæ°å¦â¦
2019年度ã®ã»ã両ãªã¼ã°ã®ãããã£ã¼ã®æ績ãã¼ã¿ãåç¸é¢ä¿æ°ã§ã¿ã 2019年度ã®ã»ã両ãªã¼ã°ã®ãããã£ã¼ã®æ績ãã¼ã¿ã®å¤æ°éã®åç¸é¢ä¿æ°ãã©ããªã£ã¦ãããè¦ã¦ã¿ãï¼ ç¹ã«åå©æ°ã¨é²å¾¡çã«çç®ããï¼ repository: github.com MathJax.Hub.Config({ tex2â¦
pythonã§åç¸é¢ä¿æ°è¡å(pcor)ãè¨ç® 以åï¼3å¤æ°(X,Y,Z)ã®å ´åã®åç¸é¢ä¿æ°ã®å¼ãå°åºããï¼ cartman0.hatenablog.com ä»åã¯ï¼å¤æ°ä»¥ä¸ã®å¤å¤æ°ã®å ´åã®åç¸é¢ä¿æ°è¡åãæ±ãã¦ã¿ãï¼ ãã scipy.stats, pandas, statsmodels,scikit-learnãªã©ã§é¢æ°ããªãâ¦
T種é¡ä¸ã®ãã®ãx種é¡å¼ãã®ã«è©¦è¡åæ°nåããã確ç ååã¯ï¼T種é¡ä¸ã®ãã®ãnåå¼ãã¦x種é¡å½ã¦ã確çï¼nï¼åºå®å¤ï¼xï¼ç¢ºçå¤æ°ï¼ãæ±ãããï¼ cartman0.hatenablog.com ãããå©ç¨ãã¦T種é¡ä¸ã®ãã®ãx種é¡å¼ãã®ã«è©¦è¡åæ°nåããã確ç$P(n|x)$ï¼ã¤ã¾ãâ¦
T種é¡ä¸ã®ãã®ãnåå¼ãã¦x種é¡å½ã¦ã確ç ä»åã¯ï¼T種é¡ä¸ã®ãã®ã nåå¼ãã¦x種é¡å½ã¦ã確ç ãæ±ããï¼ (試è¡åæ°nã¯åºå®å¤ï¼x種é¡ã確çå¤æ°ã«ãªã) ä¾ãã°ï¼3種é¡ã®ç©å ·ä»ããèåãã£ãã¨ãã«ï¼ 4åå¼ãã¦1,2,3種é¡å½ã¦ã確ç ã«è©²å½ããï¼ ã¤ã¡ã¼ã¸ Mâ¦
xåç®ã«åãã¦å½ãã確çï¼å¹¾ä½(first sucess, geometric)åå¸ã¡ã¢ xåç®ã«åãã¦å½ãã確çåå¸ã§ããå¹¾ä½(first sucess, geometric)åå¸ã®ã¡ã¢ MathJax.Hub.Config({ tex2jax: { inlineMath: [['$','$'], ['\\(','\\)']], displayMath: [ ['$$','$$'], ["\â¦
ã°ã©ãã£ã«ã«ã¢ããªã³ã°ã»æ©æ¢°å¦ç¿ãªã©ã§ãã使ãæ£è¦åå¸ã®æ§è³ª ã°ã©ãã£ã«ã«ã¢ããªã³ã°ã»æ©æ¢°å¦ç¿ãªã©ã§ãã使ãæ£è¦åå¸ã®æ§è³ªã®ã¡ã¢ éæ追å äºå®ï¼ åèï¼ SimonJ.D.Prince,Computer Vision Models, Learning, and Inference MathJax.Hub.Config({ tex2â¦
EMã¢ã«ã´ãªãºã ã§æ··åæ£è¦åå¸(MoG)ã®ãã©ã¡ã¼ã¿ãå°åº EMã¢ã«ã´ãªãºã ã§æ··åæ£è¦åå¸(MoG)ã®ãã©ã¡ã¼ã¿ãæ±ããã¡ã¢ï¼ åèï¼ Udemyã®ããã¤ãºæ¨å®ã¨ã°ã©ãã£ã«ã«ã¢ãã«ï¼ã³ã³ãã¥ã¼ã¿ãã¸ã§ã³åºç¤1 54. Gaussian Mixture Model æ··åæ£è¦åå¸ã ããã¹ãï¼"â¦
ãã¯ãã«ã»è¡åã®å¾®åã¡ã¢ 確çã»çµ±è¨ã»ã°ã©ãã£ã«ã«ã¢ããªã³ã°ã»æ©æ¢°å¦ç¿ãããã«åºã¦ããããªãã¯ãã«ã»è¡åã®å¾®åã®ã¡ã¢ éæ追å äºå® åèï¼ https://www.ics.uci.edu/~welling/teaching/KernelsICS273B/MatrixCookBook.pdf MathJax.Hub.Config({ tex2jaâ¦
å¤æ¬¡å æ£è¦åå¸ã®å¨è¾ºåå¸ã¨æ¡ä»¶ä»åå¸ãè¨ç®ãã å¤æ¬¡å æ£è¦åå¸ã®å¨è¾ºåå¸ã¨æ¡ä»¶ä»åå¸ãã¾ãæ£è¦åå¸ã«ãªãï¼ ãã®ã¨ãã®ãã©ã¡ã¼ã¿ï¼å¹³åï¼åæ£ï¼ãå°åºããï¼ D次å æ£è¦åå¸ã«å¾ã確çå¤æ°ãã¯ãã«ã$\vec{x}$, å¹³åãã¯ãã«ã$\vec{\mu}$, å ±åæ£è¡åâ¦
2ã¤ã®æ£è¦åå¸ã®å¯åº¦(pdf)ã®ç©ããå°åºã§ããæ£è¦åå¸ æ£è¦åå¸ã®ç©ãã¾ãæ£è¦åå¸ã«ãªãã®ã§ï¼ãã®æ£è¦åå¸ã®ãã©ã¡ã¼ã¿ï¼å¹³åï¼åæ£ï¼ãå°åºããï¼ ï¼ãªãï¼ç¢ºçå¤æ°ã®ç©ã§ã¯ãªãï¼ åèï¼ Udemyã®ããã¤ãºæ¨å®ã¨ã°ã©ãã£ã«ã«ã¢ãã«ï¼ã³ã³ãã¥ã¼ã¿ãã¸ã§ã³â¦
æ£è¦éã¬ã³ãåå¸ã®äºå¾åå¸ã®ãã©ã¡ã¼ã¿æ´æ°å¼ã¨äºæ¸¬åå¸ã®å°åº Udemyã®ããã¤ãºæ¨å®ã¨ã°ã©ãã£ã«ã«ã¢ãã«ï¼ã³ã³ãã¥ã¼ã¿ãã¸ã§ã³åºç¤1ãï¼ããã¹ãï¼"Computer vision: models, learning and inference" by Simon Princeï¼ã§ï¼ äºå¾åå¸ã®ãã©ã¡ã¼ã¿å¼ã®ã¿â¦
æ£è¦éã¬ã³ãåå¸ã®ç¢ºçå¯åº¦é¢æ°ã®å°åºã»å¯è¦åã»ãµã³ããªã³ã° Udemyã®ããã¤ãºæ¨å®ã¨ã°ã©ãã£ã«ã«ã¢ãã«ï¼ã³ã³ãã¥ã¼ã¿ãã¸ã§ã³åºç¤1ãã®ææ¥ã§ï¼ ãã¤ãºçµ±è¨ã§æ£è¦åå¸ã®ãã©ã¡ã¼ã¿ã®åå¸ã«ä½¿ãããæ£è¦éã¬ã³ãåå¸ã®ç´¹ä»ããã£ãã®ã§ï¼å°åºã¨å¯è¦åã®ã¡â¦
æ£è¦éã¬ã³ãåå¸ã®åæ£ã«ã¤ãã¦ã®ç©åããtåå¸ãå°åº MathJax.Hub.Config({ tex2jax: { inlineMath: [['$','$'], ['\\(','\\)']], displayMath: [ ['$$','$$'], ["\\[","\\]"] ] } }); è£é¡çãªå 容ï¼ä½ã¨ãå°åºã§ããã®ã§ã¡ã¢ï¼ ãã®è£é¡ã¯ï¼ãã¤ãºçµ±è¨ã§â¦
k-meansã¯ã©ã¹ã¿ãªã³ã°å®è£ ã¡ã¢ k-meansã¯ã©ã¹ã¿ãªã³ã°ãå®è£ ãã¦ã¿ãã®ã§ã¡ã¢ï¼ gist: k-means clustering · GitHub k-meansã¯ã©ã¹ã¿ãªã³ã°å®è£ ã¡ã¢ k-meansã¯ã©ã¹ã¿ãªã³ã°ã®ã¢ã«ã´ãªãºã å®è£ ã»ã³ã¼ã test 2次å æ¨æºæ£è¦åå¸ãã¯ã©ã¹ã¿ãªã³ã° åèæç®ã»â¦
åç¸é¢ä¿æ°ã®å°åºã¡ã¢ github: https://github.com/Cartman0/MultivariateAnalysis/blob/master/PartialCorrelationCoefficient_%E5%81%8F%E7%9B%B8%E9%96%A2%E4%BF%82%E6%95%B0%E3%81%AE%E5%B0%8E%E5%87%BA.ipynb åç¸é¢ä¿æ°ã®å°åºæ¹æ³ã®ã¡ã¢ï¼ åèè³æï¼ â¦
ABOéºä¼ååã®æ¬¡ä¸ä»£åå¸ããã«ã³ãæ¨ç§»ã§ã¿ã ABOéºä¼åå(AA, AO, BB, BO, AB, OO)ã®æ¬¡ä¸ä»£åå¸ããã«ã³ãæ¨ç§»ããã¦ï¼ ãã«ã³ãå®å¸¸ããã®ãçºæ£ããã®ããã·ãã¥ã¬ã¼ã·ã§ã³ãã¦ã¿ãï¼ ãã£ããã¨ãã¦ã¯ï¼æ¥æ¬ã ã¨BBåã®äººã¯3%ããããªãï¼åç´ã«ä¸ä»£ãç¹°â¦
cosé¡ä¼¼åº¦ã®æ¬¡å ã®åªã å ãã¿ã¯ãã¡ãã®è¨äº ã³ãµã¤ã³é¡ä¼¼åº¦ãé«ããã¯ãã«ã¯ã©ããããä¼¼ã¦ããã(岩波ãã¼ã¿ãµã¤ã¨ã³ã¹åè¡ã¤ãã³ããã) - æ¨æä¸è¶³ï¼ cosé¡ä¼¼åº¦ã¯è¨ç®ããããã®ã§ï¼è¨èªå¦ççéã§ã¯åèªãã¯ãã«ãç»åçå¦ççéã§ã¯ãã¹ãã°ã©ã ããâ¦
2ã¤ã®ãã¹ãã°ã©ã ããKLæ å ±éãè¨ç®ãã MathJax.Hub.Config({ tex2jax: { inlineMath: [['$','$'], ['\\(','\\)']], displayMath: [ ['$$','$$'], ["\\[","\\]"] ] } }); 2ã¤ã®ç¢ºçå¯åº¦é¢æ°ããKLæ å ±éãè¨ç®ã§ãããªãï¼ ï¼çµé¨çï¼ãã¹ãã°ã©ã å¯åº¦ãâ¦
é誤差ä¼ææ³ï¼ããã¯ãããã²ã¼ã·ã§ã³ï¼ã§é¢æ°ã®å¾®åå¤ãæ±ãã MathJax.Hub.Config({ tex2jax: { inlineMath: [['$','$'], ['\\(','\\)']], displayMath: [ ['$$','$$'], ["\\[","\\]"] ] } }); TheanoãPytorchã®forwardé¢æ°ãbackwardé¢æ°ãã©ãããè¨ç®â¦
æ£è¦åå¸éã®KLæ å ±éãè¨ç®ãã MathJax.Hub.Config({ tex2jax: { inlineMath: [['$','$'], ['\\(','\\)']], displayMath: [ ['$$','$$'], ["\\[","\\]"] ] } }); gist: æ£è¦åå¸éã®KLæ å ±é · GitHub KLæ å ±éã1ã2ã¨ãã£ãã¨ãã«ï¼ã©ã®ãããã®å¤§ãããâ¦
ãã£ã¡ãã«ãªããããæåã¨çµ±è¨çç¬ç«ã®ã¡ã¢ äºè±¡ã®æåã¨çµ±è¨çç¬ç«ï¼ãã®2ã¤ã¯ç¢ºçã®æç§æ¸ã®æåã®æ¹ã«ç»å ´ãï¼ èªã¿é²ãã¦ããã¨ï¼æç§æ¸ã®åé¡ã¯åºæ¬çã«æ´åããã¦ããã»ãããã¯æé»çã«ä»®å®ããã¦ããå ´åãå¤ãï¼ãããæèãã¦ããªãã¦ããã©ãã«â¦
.gist iframe.render-viewer {min-height: 1250rem} ã¬ã³ãåå¸ã¡ã¢ ãã¤ãºçµ±è¨ã«ããã¦ï¼æ£è¦åå¸ã®åæ£ãã©ã¡ã¼ã¿ã®éæ°ï¼ç²¾åº¦ï¼ãå¾ãã¬ã³ãåå¸ã«ã¤ãã¦ã®ã¡ã¢ï¼ (Jupyter Notebookã«ã¾ã¨ãã) nbviewer ã¬ã³ãåå¸ã¡ã¢ note åèæç®
Law of total Expectationsï¼å ¨æå¾ å¤åï¼ã¡ã¢ 確çã«åºã¦ããtotal Expectationsåï¼æ¥æ¬èªã ã¨å ¨æå¾ å¤åï¼ï¼ã®ã¡ã¢ï¼ 次ã®ããã«æå¾ å¤(expected) operatorã2ã¤åºã¦ããï¼ \begin{eqnarray} E[X] = E[E[X|Y]] = \sum_{y}{ E[X|Y=y]P(Y=y) } \end{eqnarraâ¦
// ã5%ã®ç¢ºçã§æ§å¨ãé²åºãããã©ããããã¯æ¬å½ã«5%ã ã£ãã®ã ã¯ã¦ããè¦ã¦ããã¨ããï¼é¢ç½ãè¨äºãè¦ã¤ãã¾ããï¼ bok.hatenablog.com ã©ãããï¼ä»¥ä¸ã®ãããªBOTã ã£ãããã§ãï¼ ã5%ã®ç¢ºçã§æ§å¨ãé²åºãããã©ããããã¨ã¯ãäºæéã«ä¸åã©ã³ãã â¦