From 7c7c99bfae53d80f78e8ba18b3dd40a112dd67e2 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Alexander=20M=C3=A4rz?= Date: Wed, 30 Aug 2023 10:41:03 +0200 Subject: [PATCH] Update example --- .../GaussianMixture_Regression_CaliforniaHousing.ipynb | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/docs/examples/GaussianMixture_Regression_CaliforniaHousing.ipynb b/docs/examples/GaussianMixture_Regression_CaliforniaHousing.ipynb index ffb44c3..7c71a77 100644 --- a/docs/examples/GaussianMixture_Regression_CaliforniaHousing.ipynb +++ b/docs/examples/GaussianMixture_Regression_CaliforniaHousing.ipynb @@ -119,8 +119,7 @@ "source": [ "# Distribution Selection\n", "\n", - "In the following, we specify a list of candidate distributions. The function dist_select returns the negative log-likelihood of each distribution for the target variable. The distribution with the lowest negative log-likelihood is selected. The function also plots the density of the target variable and the fitted density, using the best suitable distribution among the specified ones. However, note that choosing the best performing mixture-distribution based solely on training data may lead to overfitting, since mixture-densities can approximate any distribution arbitrarily well. It is therefore crucial to carefully select the specifications to strike a balance between model complexity and generalization abilit.\n", - ".\n" + "In the following, we specify a list of candidate distributions. The function `dist_select` returns the negative log-likelihood of each distribution for the target variable. The distribution with the lowest negative log-likelihood is selected. The function also plots the density of the target variable and the fitted density, using the best suitable distribution among the specified ones. However, note that choosing the best performing mixture-distribution based solely on training data may lead to overfitting, since mixture-densities can approximate any distribution arbitrarily well. It is therefore crucial to carefully select the specifications to strike a balance between model complexity and generalization ability." ] }, {