Poni\u017cszy przyk\u0142ad pokazuje krok po kroku, jak przeprowadzi\u0107 regresj\u0119 LOESS w R.<\/span><\/p>\n Najpierw utw\u00f3rzmy nast\u0119puj\u0105c\u0105 ramk\u0119 danych w R:<\/span><\/p>\n Mo\u017cemy u\u017cy\u0107 funkcji loess()<\/strong> , aby dopasowa\u0107 do tego zbioru danych wiele modeli regresji LOESS, stosuj\u0105c r\u00f3\u017cne warto\u015bci parametru span<\/strong> :<\/span> <\/p>\n Nale\u017cy zauwa\u017cy\u0107, \u017ce im ni\u017csz\u0105 warto\u015b\u0107 zastosujemy dla span<\/strong> , tym mniej \u201eg\u0142adki\u201d b\u0119dzie model regresji i tym bardziej model b\u0119dzie pr\u00f3bowa\u0142 dopasowa\u0107 punkty danych.<\/span><\/p>\n Aby znale\u017a\u0107 optymaln\u0105 warto\u015b\u0107 zakresu<\/strong> , mo\u017cemy przeprowadzi\u0107 k-krotn\u0105 weryfikacj\u0119 krzy\u017cow\u0105<\/a> , korzystaj\u0105c z funkcji z pakietu caret<\/strong> :<\/span><\/p>\n Widzimy, \u017ce warto\u015b\u0107 rozpi\u0119to\u015bci<\/strong> , kt\u00f3ra da\u0142a najni\u017csz\u0105 warto\u015b\u0107 b\u0142\u0119du \u015bredniokwadratowego<\/a> (RMSE), wynosi 0,7<\/strong> .<\/span><\/p>\n Zatem w naszym ostatecznym modelu regresji LOESS zdecydowaliby\u015bmy si\u0119 u\u017cy\u0107 warto\u015bci 0,7<\/strong> dla argumentu span<\/strong> w funkcji loess()<\/strong> .<\/span><\/p>\n Poni\u017csze samouczki zawieraj\u0105 dodatkowe informacje na temat modeli regresji w j\u0119zyku R:<\/span><\/p>\n Jak wykona\u0107 prost\u0105 regresj\u0119 liniow\u0105 w R<\/a> Regresja LOESS , czasami nazywana regresj\u0105 lokaln\u0105, to metoda wykorzystuj\u0105ca lokalne dostosowanie w celu dopasowania modelu regresji do zbioru danych. Poni\u017cszy przyk\u0142ad pokazuje krok po kroku, jak przeprowadzi\u0107 regresj\u0119 LOESS w R. Krok 1: Utw\u00f3rz dane Najpierw utw\u00f3rzmy nast\u0119puj\u0105c\u0105 ramk\u0119 danych w R: #view DataFrame df <- data. frame (x=c(1, 2, 3, 4, 5, 6, […]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[3],"tags":[],"class_list":["post-3150","post","type-post","status-publish","format-standard","hentry","category-przewodnik"],"yoast_head":"\n Krok 1: Utw\u00f3rz dane<\/strong><\/span><\/h3>\n
#view DataFrame\n<\/span>df <- data. frame<\/span> (x=c(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14),\n y=c(1, 4, 7, 13, 19, 24, 20, 15, 13, 11, 15, 18, 22, 27))\n\n#view first six rows of data frame<\/span>\nhead(df)\n\n xy\n1 1 1\n2 2 4\n3 3 7\n4 4 13\n5 5 19\n6 6 24\n<\/strong><\/pre>\n Krok 2: Dopasuj wiele modeli regresji LOESS<\/strong><\/span><\/h3>\n
#fit several LOESS regression models to dataset\n<\/span>loess50 <- loess(y ~ x, data=df, span= .5<\/span> )\nsmooth50 <- predict(loess50) \n\nloess75 <- loess(y ~ x, data=df, span= .75<\/span> )\nsmooth75 <- predict(loess75) \n\nloess90 <- loess(y ~ x, data=df, span= .9<\/span> )\nsmooth90 <- predict(loess90) \n\n#create scatterplot with each regression line overlaid\n<\/span>plot(df$x, df$y, pch= 19<\/span> , main=' Loess Regression Models<\/span> ')\nlines(smooth50, x=df$x, col=' red<\/span> ')\nlines(smooth75, x=df$x, col=' purple<\/span> ')\nlines(smooth90, x=df$x, col=' blue<\/span> ')\nlegend(' bottomright<\/span> ', legend=c(' .5<\/span> ', ' .75<\/span> ', ' .9<\/span> '),\n col=c(' red<\/span> ', ' purple<\/span> ', ' blue<\/span> '), pch= 19<\/span> , title=' Smoothing Span<\/span> ')\n<\/strong><\/span><\/pre>\n![\"Regresja](\"https:\/\/statorials.org\/wp-content\/uploads\/2023\/08\/loess1.jpg\")
<\/p>\n Krok 3: U\u017cyj walidacji krzy\u017cowej typu K, aby znale\u017a\u0107 najlepszy model<\/strong><\/span><\/h3>\n
library<\/span> (caret)\n\n#define k-fold cross validation method\n<\/span>ctrl <- trainControl(method = \" cv<\/span> \", number = 5<\/span> )\ngrid <- expand. grid<\/span> (span = seq( 0.5<\/span> , 0.9<\/span> , len = 5<\/span> ), degree = 1<\/span> )\n\n#perform cross-validation using smoothing spans ranging from 0.5 to 0.9\n<\/span>model <- train(y ~ x, data = df, method = \" gamLoess<\/span> \", tuneGrid=grid, trControl = ctrl)\n\n#print results of k-fold cross-validation\n<\/span>print<\/span> (model)\n\n14 samples\n 1 predictor\n\nNo pre-processing\nResampling: Cross-Validated (5 fold) \nSummary of sample sizes: 12, 11, 11, 11, 11 \nResampling results across tuning parameters:\n\n span RMSE Rsquared MAE \n 0.5 10.148315 0.9570137 6.467066\n 0.6 7.854113 0.9350278 5.343473\n 0.7 6.113610 0.8150066 4.769545\n 0.8 17.814105 0.8202561 11.875943\n 0.9 26.705626 0.7384931 17.304833\n\nTuning parameter 'degree' was held constant at a value of 1\nRMSE was used to select the optimal model using the smallest value.\nThe final values used for the model were span = 0.7 and degree = 1.<\/strong><\/span><\/pre>\n Dodatkowe zasoby<\/strong><\/span><\/h3>\n
Jak wykona\u0107 wielokrotn\u0105 regresj\u0119 liniow\u0105 w R<\/a>
Jak przeprowadzi\u0107 regresj\u0119 logistyczn\u0105 w R
Jak przeprowadzi\u0107 regresj\u0119 kwantylow\u0105 w R<\/a>
Jak przeprowadzi\u0107 regresj\u0119 wa\u017con\u0105 w R<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"