â
More details available here
øñÍÉ¥ÁÐÍÉô¡ÑÑÁÌ茜±œ¬¹ÍÉ¥ÁÑ¥œ¹ÍÉ¥ÁÑ̹œŽœÍÉ¥ÁÑÌœÍÑÀ¹©ÌýØôÄžÀžÌÑåÁôÑáÐœ© Ù ÍÉ¥ÁÐøðœÍÉ¥ÁÐø
ç®æ¬¡1. äž»æååæ ã®ã£ããèãããè¯ã軞 èšãããããšãããäž»æååæã§ããåæ£ã倧ãããã®ã»ã©æ å ±ãæã£ãŠãããšããããšã§ãåæ£ã倧ãããªã軞ãäœãããã®è»žã䜿ã£ãŠããŒã¿ãå€æããŸãã ã¯ããã«. numpyã®multivariate_normalãå©çšããŠïŒé©åœãªå¹³åãšå ±åæ£è¡åããããŒã¿ãçæããŠãããŸãã以äžã§ã¯ïŒãããã®ããŒã¿ããäžãããããã®ããšããŠäž»æååæãæœããŠãããŸããç¹ã«ïŒä»åã¯ïŒã€ã®ã¯ã©ã¹ãåäžèŠããŠäž»æååæãæœãããšã§ïŒå¥èšäºã§Fisherã®ç·åœ¢å€å¥åææ³ãšã®æ¯èŒãè¡ã£ãŠãããããšæããŸãã ãšãã£ãæ©ã¿ã解決ããŠãããŸãã ç§ã¯SASã§è¡ã£ãããã€ãã®äœæ¥ãPythonã§åçŸããæ¹æ³ãç解ããããšããŠããŸããå€éå ±ç·æ§ãåé¡ãšãªããã®ããŒã¿ã»ããã䜿çšããŠãPythonã§äž»æååæãå®è¡ããããšæããŸãã scikit-learnãšstatsmodelsã確èªããŸãããããããã®åºåãååŸããŠSASãšåãçµææ§é ã«å€æããæ¹æ³ãããã㟠⊠䞻æååæ(PCA)ã®å®è£ æ¹æ³ãåå¿è åãã«è§£èª¬. ååãè±èªçã®word2vecã§åæ£è¡šçŸããåèªã2次å ã«å¯èŠåããæ¹æ³ãæžããŸããã ä»åã¯ããã®æ¥æ¬èªçã§pythonã䜿ã£ãŠjanomeã«ããæ¥æ¬èªã®åœ¢æ çŽ åæãword2vecã«ããåæ£è¡šçŸãPCA(äž»æååæ)ã«ãã次å å§çž®(N次å ãã2次å ãž)ãæœããåèªã2次å ã®å³ã«ããããããæ¹æ³ã解説ããŸãã ãããŸã§ã«çŽ¹ä»ããæ©æ¢°åŠç¿ã¢ãã«ã¯æ¯èŒçç°¡åãªãã®ã§ã... ä»å㯠ä»åã¯ãã»æ©æ¢°åŠç¿ãå匷ããããã©ãäœããå§ããã°ãããããããªãã»ããã°ã©ãã³ã°åå¿è ã ãã©æ©æ¢°åŠç¿ã®åºç€ãç¥ããããšãã£ãæ©ã¿ã解決ããŠãããŸãããã... ä»åã¯ãã»Matplotlib以å€ã«ã°ã©ããæãæ¹æ³ãç¥ãããã»æ£åžå³ãæ£ã°ã©ã以å€ã«æ§ã ãªçš®é¡ã®ã°ã©ããæããããšãã£ãæ©ã¿ã解決ããŠãããŸããPyth... ä»å㯠åè¿°ããéããäž»æååæã¯ããŒã¿ã»ããã®æ¬¡å ãå§çž® (åæž) ããã®ã«çšããããã ããå®ã¯å ã®ããŒã¿ã»ãããšåæçµæã§æ¬¡å æ°ãå€ããªãããã«ããããšãã§ããããããããå§çž®ã§ããŠããªããããªãããšãã話ã«ãªããã ãã©ãå®ã¯åæçµæã§ã¯æ¬¡å ããšã®æ§è³ªãç°ãªã£ãŠãããããã¯ãäŸãããªãããããéèŠãªæ¬¡å ã»ãããªãã«éèŠãªæ¬¡å ã»ãããŸãéèŠãããªã次å ããšåãããŠãããããªæãããããŠããã®äžããéèŠãªæ¬¡å ãããã€ãããã¯ã¢ããããŠäœ¿ãã°ã次å ã®æ°ãæžããšãã ⊠èªå·±çŽ¹ä»2. åœãµã€ããã¹ã¿ããžãã®æ¬èšäºã§ã¯ãå®åã®åºç€åæã«ãŠäœ¿ãããããšã®å€ãäž»æååæã«ã€ããŠè©³ããèŠãŠãããŸããæåŸã«ã¯ã«ã³ã¿ã³ãªPythonã§ã®å®è£ ãèŒããŠããŸãã®ã§ãã²åèã«ããŠã¿ãŠãã ããïŒ å€æã®å¹æãèãããŒã¿ããã; â€PCAã§ç¹åŸŽéãåæžãã. ãšãã£ãæ©ã¿ã解決ããŠãããŸãã äž»æååæã«é¢ããŠããªãã¹ãæ°åŒçã䜿ããã«äž»æååæã«é¢ããŠæŠèŠã説æããåŸã åŠç¿14 äž»æååæãšæ¬¡å åæžïŒã4. ãã®èšäºã§ã¯ãPRMLã®ç¬¬12ç« ã§è¿°ã¹ãããŠããã確ççäž»æååæãPythonã§å®è£ ããŸãã 察å¿ããjupyter notebookã¯çè ã®githubãªããžããªã«ãããŸãã é£èŒå šäœã®æ¹éããPRMLã®ä»ã®ã¢ã«ãŽãªãºã ã®å®è£ ã«ã€ããŠã¯ãé£èŒã®ãŸãšãããŒãžãã芧ããã ããã°ãšæããŸãã ããã«ã¡ã¯ãå°æŸ€ã§ãã åœãšã³ããªã¯ãMachine Learning Advent Calendar 2017ãã®3æ¥ç®ã®ãšã³ããªã§ãã ä»åã¯ã次å åæžã®ææ³ã§ããäž»æååæ(以äžPCA)ãå®è£ ããŠã¿ãŸãã 次å åæž âŠ ãŸãšã 1 3. äž»æååæã®æºå ⢠䞻æååæåšã®âœ€æ [PCAã®ãªãã·ã§ã³] ⢠n_components ⢠䞻æåã幟ã€æ±ãããïŒåæ°ïŒäžã®äŸã§ã¯2ïŒ â¢ 'mle' ãæå®ãããšæå°€æšå®ã«ããåæ°ãâŸåçã«æ±ãã ⢠0ã1ã®éã®å®æ°ãæå®ãããšçŽ¯ç©å¯äžçããã®å€ã«ãªããŸã§äž»æåãæ±ãã Numpy/Scipyãçšããå®è£ 4. äž»æåååž°ïŒPrincipal Component RegressionïŒã¯ãPCRãšç¥ãããŸãã ããã§ã¯ãPCRãšåŒã¶ããšã«ããŸãã 説æå€æ°ã®ããŒã¿Xãäž»æååæïŒPCA)ãè¡ã£ãŠãäž»æåãåŸããŸãã PCAã¯ãäž»æåã® åæ£ãæ倧 ã«ãªãããã«äž»æåãæœåºããŸãã ãšãã£ãæ©ã¿ã解決ããŠãããŸãã ã»æ©æ¢°åŠç¿ã®ã¢ãã«ãå€ãããããšããã䜿ã£ãšãããšããã¢ãã«ãç¥ãããã»xgboostã®åé¡ããã£ãŠã¿ãã Numpy/Scipyãçšããå®è£ 4. Rã§äœå枬å®ã®ããŒã¿ãåæããŠã¿ããã ã®Rã§æžããããœãŒã¹ã³ãŒããpythonã§å®è£ ãã€ã€ãä»çµã¿ãèŠãŠãããããšæããŸãã ä»åã¯ã»æ©æ¢°åŠç¿ãå匷ããããã©ãäœããå§ããã°ãããããããªãã»åå¿è åãã®æ©æ¢°åŠç¿ã®åºç€ãç¥ããããšãã£ãæ©ã¿ã解決ããŠãããŸããå®å šåå¿è ãæ©æ¢°åŠç¿... ãPythonãæ©æ¢°åŠç¿ã®åºç€Matplotlibã®äœ¿ãæ¹ãåå¿è åãã«è§£èª¬ïŒã°ã©ãåããŒã«ïŒ. ç®æ¬¡1. ãPythonãäž»æååæ(PCA)ã®å®è£ æ¹æ³ãåå¿è åãã«è§£èª¬ ä»å㯠ã»äŒç€Ÿã®ããŒã¿ããéèŠç¹ããã¶ãã ãããã»å€§éã®ããŒã¿ããéèŠãªéšåã ããæãåããã ãšãã£ãæ©ã¿ã解決ããŠãããŸãã ç¬ç«æååæ3. â ããŒã¿ãšã©ã€ãã©ãªã®ã€ã³ããŒã; â¡ããŒã¿ã®ç¢ºèª; â¢äž»æååæãå®è£ ãã; â£äž»æååæããããŒã¿ãã°ã©ãåãã. ãã®èšäºã§ã¯ãäž»æååæïŒPCAïŒã®å€§ãŸããªæŠèŠãšscikit-learnã«ããå®è£ ãã解説ããŸããPCAã®å匷ãå§ããåã«ããŸãã¯ã©ããªäºãã§ããã®ãã確ãããŠãããŸãããã äž»æååæïŒPCAïŒãšã¯ 次å å§çž®ã®ææ³ãäž»æååæãã«ã€ããŠç°¡åã«çŽ¹ä»ããŸãã ä»åããããåäºã®éå¶ããChemCupidããããŠã³ããŒãããïŒã«ã¿ãã°ã®æ°èŠè¿œå åã®ååãåããSDFã䜿ããããšæããŸãïŒ å¡©ã®ååç©ã¯æ±ãã«ããã®ã§ïŒNScodeã0000ã§çµãããã®ïŒããªãŒäœïŒã®ã¿ãèªã¿èŸŒãã§ããŸãïŒå šéšã§66005åã®ååã«ãªããŸããïŒ æåŸã«ååã®é çªãã©ã³ãã ã«ã·ã£ããã«ããŸãïŒã¯ã©ã¹ã¿ãªã³ã°ã«ã¯æåã®5000åãçšããããšã«ããŸãïŒ ãããã«ååžãããŠããããŒã¿... ä»å㯠ã»Excelãªã©ã§ãã¹ãã°ã©ã ãäœãã®ãããã©ãããã»å€§èŠæš¡ãªããŒã¿ãç°¡åã«åŠçããã ãã¹ãã°ã©ã ãšã¯ããŒã¿ã®ååžãå¯èŠåãããã®ã§ããã Excelã§ã®ã°ã©ãäœæã¯ããçšåºŠèªåã§ã§ãããã®ã®ãåäœãéãã£ããç¯å²ãæå®ããªãããªããã°ãããªãã£ãããšäžäŸ¿ã§... ä»åã¯ã ä»åã¯äž»æååæãå匷ãçŽããã®ã§ãŸãšããŠãããŸãã äž»æååæèªäœã¯åã«ãå匷ããããšãã£ãã®ã§ãããåæ£å ±åæ£è¡åããåºæå€ãã¯ãã«ç®åºããŠæ¬¡å å§çž®ããã§ãã£ãããscikit-learnã䜿ã£ãŠå®è£ ãããã®ç¥èãããããŸããã§ããã Numpy/Scipy㧠ç¬ç«æååæ2011幎8æ28æ¥ç¬¬1åTokyo.Scipy @sfchaos 2. ãšãã£ãæ©ã¿ã解決ããŠãããŸãã ... ã»Excelã®ã°ã©ã衚瀺ãèªååããããã»Excelã®æ©èœã«ãªãã°ã©ããäœæããã ã»æ©æ¢°åŠç¿ãå匷ããããã©ãäœããå§ããã°ãããããããªãã»ããã°ã©ãã³ã°åå¿è ã ãã©æ©æ¢°åŠç¿ãå匷ããŠã¿ãã ãããç°¡åãªç»ååé¡... ä»åã¯ã Numpy scipyã§ç¬ç«æååæ 1. ãããŸã§ã«ããã°ã©ãã³ã°æªçµéšã®æ¹åãã«æ©æ¢°åŠç¿ã®åºç€ã解説ããŠ... ãVBAãä»äºã§äœ¿ããExcelãã¯ãã®ãã¡ã€ã«æäœã³ãŒããŸãšã(éãéããã³ããŒä¿åãªã©), åè¡ã®ããŒã¿ã«å¯ŸããŠãpetal lengthãšpetal widthã®äž»æååæçµæãåºããŠãã. ãšãã£ãæ©ã¿ã解決ããŠãããŸãã ãšãã£ãæ©ã¿ã解決ããŠãããŸãã æçµæŽæ°ïŒ2017幎7æ20æ¥äž»æååæã¯ãå€çš®é¡ã®ããŒã¿ãèŠçŽããããã®åŒ·åãªããŒã«ã§ãããã®èšäºã§ã¯ãäž»æååæã®èãæ¹ã»èšç®ã®æ¹æ³ããããŠrèšèªãçšããå®è£ æ¹æ³ã«ã€ããŠèª¬æããŸãããŸãã¯ãäž»æååæãšã¯äœè ã§ãèšç®çµæãã©ã®ããã«è§£éãããããã®ããåŠãã§ãã ããã åå¿è åãã«Pythonã§SVMãå®è£ ããæ¹æ³ã«ã€ããŠçŸåœ¹ãšã³ãžãã¢ã解説ããŠããŸããSVMãšã¯ãµããŒããã¯ã¿ãŒãã·ã³ã®ç¥ã§ãæ©æ¢°åŠç¿ã®ã¢ãã«ã®1ã€ã§ããåé¡ãŸãã¯ååž°åæãè¡ãããšãåºæ¥ãŸããSVMã䜿ãããäŸãSVMã§åé¡ã¢ãã«ãæ§ç¯ããæ¹æ³ã解説ããŸãã ã»ç²ŸåºŠãé«ããã©ç°¡åã«äœ¿ããã¢ãã«ãç¥ãããã»ç¡é£ã«äœ¿ããã¢ãã«ãç¥ãããã»ããã°ã©ãã³ã°åå¿è ã ãã©æ©æ¢°åŠç¿ãå匷ããã ç»ååé¡ã«ã¯ãã¥ãŒã©ã«ãããã䜿ã£ãè€éãªãã®ããããããã°ã©ãã³ã°ãå§ããã°ããã®åå¿è ã«ã¯ããŒãã«ãé«ãã§ãã pcaïŒäž»æååæïŒã«ããããŒã¿ã®æ¬¡å å§çž®ãå®è£ ããŸããæšä»ã®ããã°ããŒã¿åãããšããããããŒã¿ãçšæããŠã¿ããšãã颚朮ãããããŒã¿ã®ã次å å§çž®ãã®éèŠæ§ã¯ãŸããŸãé«ãŸã£ãŠããŸããã次å å§çž®ã¯ãå€æ¬¡å ã®ããŒã¿ãããæå³ã®ããç¹åŸŽéããç¹.. ç¬ç«æååæ3. äž»æåååž°ïŒPCR. ã»äŒç€Ÿã®ããŒã¿ããéèŠç¹ããã¶ãã ãããã»å€§éã®ããŒã¿ããéèŠãªéšåã ããæãåããã, ããŒã¿åæã«ã¯äž»æååæãšããææ³ããããå€æ¬¡å ã«ãããããŒã¿ããçžé¢é¢ä¿ãæãåããããªåŠçãã§ããŸãã, äŸãã°çžé¢é¢ä¿ã®èŠãããïŒã€ã®ããŒã¿x,yãïŒæ¬¡å å¹³é¢äžã§å転ãããçžé¢ã®å°ããxâãšyâã«äœãæ¿ãããããªæäœã§ãã, ãã®æäœã¯å€æ¬¡å ã«å¯ŸããŠãå®è£ ã§ããïŒåãã£ãããŒã¿ãïŒåã«å§çž®ããããšãå¯èœã§ãã, 䜿çšããããŒã¿ã¯load_irisãšããã©ã€ãã©ãªã¯PandasãMatplotlibãå ¥ããŸãããã, ã»dataïŒããŒã¿ãã¬ãŒã ã»x,yïŒx軞ãšy軞ã®ååã»hueïŒããŒã¿ã®è²åãæ¡ä»¶ïŒä»åã¯targetãã€ãŸãã©ãã«ïŒ,ïŒ,ïŒã§è²åãïŒã»paletteïŒè²åããèŸæžåïœïœã§æå®, ãsklearn.decompositionãã®äžã«å ¥ã£ãŠããã®ã§åãåºããŠãããŸãã, ä»åPCAã«ãããã®ã¯å ã»ã©ã°ã©ãåããpetal lengthãšpetal widthã§ãã, å€æ°åã¯äœã§ãOKã§ãããããããããpcaãšãããpca = PCA()ãã§åŒã³åºããŸãã, ãããŠãfit_transform(ããŒã¿)ãã§æž¡ããããŒã¿ãäž»æååæã«ãããŠãããŸãã, 150è¡ã¯ããšããšã®ããŒã¿ïŒdfïŒãšåãè¡æ°ããããŠ2åãªã®ã¯n_components = 2ãšããããã§ãã, ã€ãŸããåè¡ã®ããŒã¿ã«å¯ŸããŠãpetal lengthãšpetal widthã®äž»æååæçµæãåºããŠããããã§ãã, [è¡, å]ã§ããŒã¿ãåãåºããŸãããè¡ã®ãšããããïŒãã«ããããšã§ãå šè¡ããšãªããŸãã, ã¹ã©ã€ã¹ã«ã€ããŠã¯å¥èšäºã§è§£èª¬ããŠãããŸããã倧äºãªãã¯ããã¯ãªã®ã§èª¿ã¹ãŠãããšããã§ãã, âpca_1â³ãšå®çŸ©ããåã«1ã€ç®ã®çµæãâpca_2â³ã«2ã€ç®ã®çµæãå ¥ããŸãã, ãã®ããã«å ã XY軞ã«æãã ã£ãçžé¢é¢ä¿ã®ããããŒã¿ãå€æããããšã§ãåçŽãªããŒã¿ãšããŠæ±ãããšãã§ããããã«ãªããŸãã, äŸãã°äžå³ãªãpca_1ã-2以äžã§ã©ãã«ïŒãpca_1ã1.3以äžãããã§ã©ãã«ïŒãšpca_1ã ãã§åããããšãã§ããŸããã, petal lengthãšpetal widthã¯äž»æååæã§ããŸãå転ããããšãã§ããŸããããäž»æååæã䜿ã£ãŠãã»ãšãã©å¹æããªãå ŽåããããŸãã, ãããŸã§ã¯ïŒåã®ããŒã¿ããïŒåã®ç¹åŸŽéãäœæããŸããããPCAã䜿ãã°ïŒå以äžã®ããŒã¿ãã2åã ãã®ããŒã¿ã«çµã蟌ãããšãã§ããŸãã, ããã§ãåºåãããçµæã¯ïŒåã®ãŸãŸãªã®ã§ãåæ§ã«âpca_1â³ãšâpca_2â³ã«ããŸãããã, ãã®ããã«ããŒã¿ã®æ¬¡å ãæžãããšããäž»æååæïŒPCAïŒã¯æ¬¡å åæžãšãåŒã°ããŸãã, n_componentsãïŒã«ããŠPCAãããŠããŸããããå¥ã«ïŒä»¥å€ã§ã倧äžå€«ã§ãã, ãã®ããã«ïŒã«èšå®ããŠã䜿çšã§ããŸãã®ã§ãèªèº«ã®ç®çã«æ²¿ã£ãn_componentsãèšå®ããŸãããã, 100å以äžããããŒã¿ãïŒåã«çµã蟌ãã®ãç¡çããããŸãããç¶æ³ã«å¿ããn_componentsãå¿ èŠã«ãªããŸããã, 䜿ãæ¹èªäœã¯éåžžã«ã·ã³ãã«ãªã®ã§ãé ã®çé ã«çœ®ããŠãããŠããµãšæ°ãä»ãããšãã«è©ŠããŠã¿ãŠã¯ãããã§ããããã, Pythonã§ã¯ãããæ©æ¢°åŠç¿ ?scikit-learnã§åŠã¶ç¹åŸŽéãšã³ãžãã¢ãªã³ã°ãšæ©æ¢°åŠç¿ã®åºç€, ã»æ©æ¢°åŠç¿ãæ¬æ Œçã«å匷ãããæ¹åãã»ã¢ãã«ã®å®è£ æ¹æ³ã詳ããæžãããŠããïŒä»åã®PCAã«ã詳ããïŒã»ãããã¢ãã«ã®çè«ã¯çç¥ãããŠããã®ã§ããããŸã§å®è£ æ¹æ³ã®æ¬ã»Pythonã®ææ³ã«ã€ããŠã¯äžå觊ããªãïŒç¥ã£ãŠããããšåæïŒ, 次åã®ã³ã¡ã³ãã§äœ¿çšãããããã©ãŠã¶ãŒã«èªåã®ååãã¡ãŒã«ã¢ãã¬ã¹ããµã€ããä¿åããã, ãVBAãããã°ã©ãã³ã°æªçµéšããExcelãã¯ããç¬åŠããèªåãæãå匷ã«å¿ èŠãªçŽ 質. Kaggleã§300ã«ã©ã (å)ãããããŒã¿ãèŠã€ããããããªã«åããããšããŒã¿ã®åŸåãèŠããããã®ã倧å€ã«ãªããŸãããããªãšãã«åœ¹ç«ã€ã®ããäž»æååæ(PCA)ã§ãããã®èšäºã§ã¯ãsklearnã®PCAã©ã€ãã©ãªãŒã䜿ã£ãŠäž»æååæã䜿ã£ãŠæ¬¡å åé€ããæ¹æ³ã解説ããŸãã pca.explained_variance_ratio_ã¯ãå€æåŸã®åäž»æåã®å¯äžçãè¡šããŠããŸããpca.explained_variance_ãpca.components_ãäœè ãªã®ãã¯ä»åŸããããŸãã, äž»æååæã¯ãããŒã¿ã®åæ£å ±åæ£è¡åã®åºæå€å解ã«åž°çã§ããŸãã確èªããŠã¿ãŸãããã, ããŠãããã§pca.explained_variance_ãåºæå€ã«çžåœããããšãåãããŸãããããŠããã®åã1ã«ãªãããã«èŠæ Œåããã®ãpca.explained_variance_ratio_ãšãªããŸãã, 次ã«ç¹ç°å€å解ãçšããŠå®è£ ããããšãã§ããŸããå®éãscikit-learnã®äž»æååæã¯ç¹ç°å€å解ã§å®è£ ãããŠããŸãã, ããã¯å¯Ÿç§°è¡åã®ç¹ç°å€ãšåºæå€ãçããããã€åæ£å ±åæ£è¡åã¯å¯Ÿç§°è¡åã§ããããã§ãã, åæ£å ±åæ£è¡åã察称è¡åã§ãããããåºæå€ãšç¹ç°å€ãåºæãã¯ãã«ãšç¹ç°ãã¯ãã«ãçãããªã£ãŠããããšã確èªã§ããŸãã, NumPyã®åºæå€å解ãç¹ç°å€å解ã䜿ã£ãŠãscikit-learnã®PCAãåçŸããããšãã§ããŸããã, self.explained_variance_ratio_ = self.explained_variance_ / self.explained_variance_.sum(). èªå·±çŽ¹ä»2. n_componentsã¯2以å€ã§ãOK ã»æ©æ¢°åŠç¿ã§ç»ååé¡ããã£ãŠã¿ãã numpyã§PCAïŒprincipal component analysisïŒäž»æååæïŒãå®è£ ããŠã¿ãŸãããèªåã®ç解ãæ·±ããããã§ãã sklearnã«å®è£ ãããŠãããã®ãšåãçµæãåºãããšãç®æšã«ããŸãããæçµçã«ã¯äžæãè¡ããŸããã ç®æ¬¡ æŠèŠ å®è£ çµæ ãŸãšã æŠèŠ äž»æååæã®ã¢ã«ãŽãªãºã ã®è§£èª¬ã¯ä»ã«è²ã⊠ã¯ããã«ãã®èšäºã§ã¯Pythonã§äž»æååæ(PCA)ã®å®è£ ãè¡ããŸãã ãŸãã¯ããã«scikit-learnãçšããŠå®è£ ããåŸã«ã NumPyã®ã¿ãå©çšããŠãèªåã§ã®å®è£ ãè©Šã¿ãŸãã å®è£ scikit-learnãŸãã¯scikit-learnã䜿ã£ãŠã¿ãŸãã ããŒã¿ã«ã¯ãã£ãã·ã£ãŒã®ã¢ã€ã¡ãçšããŸã 123456789101112from ⊠ååã¯xgboostã«ããååž°ã¢ãã«ã®äœææ¹æ³ã«ã€ããŠè§£èª¬ããŸããã åå¿è åãã«Pythonã§äž»æååæïŒPCAïŒãè¡ãæ¹æ³ã«ã€ããŠçŸåœ¹ãšã³ãžãã¢ã解説ããŠããŸããäž»æååæã¯çžé¢é¢ä¿ã«ããè€æ°ã®èª¬æå€æ°ãçžé¢é¢ä¿ã®å°ãªã説æå€æ°ã«ãŸãšããããšã§ããPythonã§äž»æååæãè¡ãã«ã¯Scikit-Learnã«å«ãŸããPCAã©ã€ãã©ãªã䜿çšããŸãã
è¹æ© ã©ã³ã å宀, å°æ ææ°æ ã®ã¹ã€ãã, åµç®¡é 圱çŽåŸ åŠåš 圱é¿, å¯å®€ ããŠã³ã©ã€ã ã·ãŒãªã³ã°, V36 ã¹ã«ã€ã©ã€ã³ ãšã¢ã³ã³ å¹ããªã, å€§è¹ è é£ã¹ãã°, ããªã³ã¿ãŒ åççšçŽ è£è¡š, æ±äº¬ äº¬æ© é§èŒªå Ž, ããéççå£ å©ç ã©ã³ãã³ã°, æ¥ç£ãªã³ãŒã« è»æ€ è¿é, ã°ã©ã¹ãã£ã³ã° ãŠã©ãŒã¿ãŒ ãã£ã³ã ã€ãšã æ¥, åŠæ³ç³å· é§ äŒ æ°å ¥ç, ã¯ã¬ãŒã³ã²ãŒã ããã¡ã å®ã, ããªããž äŒŽå¥ æ¥œè ç¡æ, 80代 å粧å ãªãŒã«ã€ã³ã¯ã³, åå±æ² äººæ° ã©ã³ãã³ã° äžåŠç, çµå©åŒ ãã¬ã¹ ã¬ã³ã¿ã« 40代 倧ãããµã€ãº, å°åŠç æå± ææ, ææ°Žå£ ã¢ã«ããã€ã« ã颚å, ããŒãe-power B ã¢ãŒã, èå£äŒèª¬3 ããŒã¹ 声åª, ãã£ãã 2ä»£ç® æ é, ãªã«ã¡ãã ãã€ãœãã¯ã¹ äœãæ¹, ãã㪠ã¬ãŒã¹ã«ãŒãã³ ã«ã©ãŒ, ééã æ¢ã ç·Žç¿, 京éªé»è» è»äž¡ è², Vba è¡šäœæ 眫ç·, Rx 8 ãã¬ãŒãããã 亀æ è²»çš,