control that predictions are > 0 using GridSearchCV
.everyoneloves__top-leaderboard:empty,.everyoneloves__mid-leaderboard:empty,.everyoneloves__bot-mid-leaderboard:empty height:90px;width:728px;box-sizing:border-box;
I am using GridSearchCV in order to estimate the parameters of my regressor.
I use the scoring function mean_squared_log_error (and I would like to continue using it)
from sklearn.model_selection import GridSearchCV
import xgboost as xgb
gs = GridSearchCV(xgb.XGBRegressor(),
param_grid='max_depth': range(5, 10),
scoring='neg_mean_squared_log_error', cv=5, return_train_score=True)
gs.fit(X, y)
y is always positive, but what happens is that some predictions within the 5 fold gridsearch happen to be negative (even if it shouldn't happen because my target variable is always positive) and therefore I get the error message
ValueError: Mean Squared Logarithmic Error cannot be used when targets contain negative values.
because the scorer is trying to calculate the log of a negative number (the unfortunate prediction).
Is there a way to control the predictions inside the GridSearchCV? How would you tackle this issue?
python scikit-learn xgboost gridsearchcv msle
asked Nov 15 '18 at 14:14
gabboshowgabboshow
1,76341847
add a comment |
I am using GridSearchCV in order to estimate the parameters of my regressor.
I use the scoring function mean_squared_log_error (and I would like to continue using it)
from sklearn.model_selection import GridSearchCV
import xgboost as xgb
gs = GridSearchCV(xgb.XGBRegressor(),
param_grid='max_depth': range(5, 10),
scoring='neg_mean_squared_log_error', cv=5, return_train_score=True)
gs.fit(X, y)
y is always positive, but what happens is that some predictions within the 5 fold gridsearch happen to be negative (even if it shouldn't happen because my target variable is always positive) and therefore I get the error message
ValueError: Mean Squared Logarithmic Error cannot be used when targets contain negative values.
because the scorer is trying to calculate the log of a negative number (the unfortunate prediction).
Is there a way to control the predictions inside the GridSearchCV? How would you tackle this issue?
python scikit-learn xgboost gridsearchcv msle
asked Nov 15 '18 at 14:14
gabboshowgabboshow
1,76341847
you could use the mean absolute error (MAE)
– makaros
Nov 15 '18 at 23:27
Or you could a use a custom wrapper which just wraps around theneg_mean_squared_log_errorand clips the negative values to 0 before passing it to the scorer, if that makes sense.
– Vivek Kumar
Nov 16 '18 at 4:39
add a comment |
I am using GridSearchCV in order to estimate the parameters of my regressor.
I use the scoring function mean_squared_log_error (and I would like to continue using it)
from sklearn.model_selection import GridSearchCV
import xgboost as xgb
gs = GridSearchCV(xgb.XGBRegressor(),
param_grid='max_depth': range(5, 10),
scoring='neg_mean_squared_log_error', cv=5, return_train_score=True)
gs.fit(X, y)
y is always positive, but what happens is that some predictions within the 5 fold gridsearch happen to be negative (even if it shouldn't happen because my target variable is always positive) and therefore I get the error message
ValueError: Mean Squared Logarithmic Error cannot be used when targets contain negative values.
because the scorer is trying to calculate the log of a negative number (the unfortunate prediction).
Is there a way to control the predictions inside the GridSearchCV? How would you tackle this issue?
python scikit-learn xgboost gridsearchcv msle
asked Nov 15 '18 at 14:14
gabboshowgabboshow
1,76341847
I am using GridSearchCV in order to estimate the parameters of my regressor.
I use the scoring function mean_squared_log_error (and I would like to continue using it)
from sklearn.model_selection import GridSearchCV
import xgboost as xgb
gs = GridSearchCV(xgb.XGBRegressor(),
param_grid='max_depth': range(5, 10),
scoring='neg_mean_squared_log_error', cv=5, return_train_score=True)
gs.fit(X, y)
y is always positive, but what happens is that some predictions within the 5 fold gridsearch happen to be negative (even if it shouldn't happen because my target variable is always positive) and therefore I get the error message
ValueError: Mean Squared Logarithmic Error cannot be used when targets contain negative values.
because the scorer is trying to calculate the log of a negative number (the unfortunate prediction).
Is there a way to control the predictions inside the GridSearchCV? How would you tackle this issue?
python scikit-learn xgboost gridsearchcv msle
python scikit-learn xgboost gridsearchcv msle
asked Nov 15 '18 at 14:14
gabboshowgabboshow
1,76341847
asked Nov 15 '18 at 14:14
gabboshowgabboshow
1,76341847
asked Nov 15 '18 at 14:14
gabboshowgabboshow
1,76341847
asked Nov 15 '18 at 14:14
gabboshowgabboshow
1,76341847
asked Nov 15 '18 at 14:14
gabboshowgabboshow
1,76341847
1,76341847
you could use the mean absolute error (MAE)
– makaros
Nov 15 '18 at 23:27
Or you could a use a custom wrapper which just wraps around theneg_mean_squared_log_errorand clips the negative values to 0 before passing it to the scorer, if that makes sense.
– Vivek Kumar
Nov 16 '18 at 4:39
add a comment |
you could use the mean absolute error (MAE)
– makaros
Nov 15 '18 at 23:27
Or you could a use a custom wrapper which just wraps around theneg_mean_squared_log_errorand clips the negative values to 0 before passing it to the scorer, if that makes sense.
– Vivek Kumar
Nov 16 '18 at 4:39
you could use the mean absolute error (MAE)
– makaros
Nov 15 '18 at 23:27
you could use the mean absolute error (MAE)
– makaros
Nov 15 '18 at 23:27
Or you could a use a custom wrapper which just wraps around the
neg_mean_squared_log_error and clips the negative values to 0 before passing it to the scorer, if that makes sense.– Vivek Kumar
Nov 16 '18 at 4:39
Or you could a use a custom wrapper which just wraps around the
neg_mean_squared_log_error and clips the negative values to 0 before passing it to the scorer, if that makes sense.– Vivek Kumar
Nov 16 '18 at 4:39
add a comment |
1 Answer
1
active
oldest
votes
If you have knowledge that your dependent (y value) is always positive you can use a loss function that constrains your predictions to be in the positive domain as well.
One example supported in XGBoost is Gamma regression (see reg:gamma) or you can design your own loss function like the Mean Squared Log Error, you'd have to derive first and second order derivatives in that case.
answered Nov 19 '18 at 14:17
BarBar
1,3011931
add a comment |
Your Answer
StackExchange.ifUsing("editor", function ()
StackExchange.using("externalEditor", function ()
StackExchange.using("snippets", function ()
StackExchange.snippets.init();
);
);
, "code-snippets");
StackExchange.ready(function()
var channelOptions =
tags: "".split(" "),
id: "1"
;
initTagRenderer("".split(" "), "".split(" "), channelOptions);
StackExchange.using("externalEditor", function()
// Have to fire editor after snippets, if snippets enabled
if (StackExchange.settings.snippets.snippetsEnabled)
StackExchange.using("snippets", function()
createEditor();
);
else
createEditor();
);
function createEditor()
StackExchange.prepareEditor(
heartbeatType: 'answer',
autoActivateHeartbeat: false,
convertImagesToLinks: true,
noModals: true,
showLowRepImageUploadWarning: true,
reputationToPostImages: 10,
bindNavPrevention: true,
postfix: "",
imageUploader:
brandingHtml: "Powered by u003ca class="icon-imgur-white" href="https://imgur.com/"u003eu003c/au003e",
contentPolicyHtml: "User contributions licensed under u003ca href="https://creativecommons.org/licenses/by-sa/3.0/"u003ecc by-sa 3.0 with attribution requiredu003c/au003e u003ca href="https://stackoverflow.com/legal/content-policy"u003e(content policy)u003c/au003e",
allowUrls: true
,
onDemand: true,
discardSelector: ".discard-answer"
,immediatelyShowMarkdownHelp:true
);
);
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
StackExchange.ready(
function ()
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fstackoverflow.com%2fquestions%2f53321395%2fcontrol-that-predictions-are-0-using-gridsearchcv%23new-answer', 'question_page');
);
Post as a guest
Required, but never shown
1 Answer
1
active
oldest
votes
1 Answer
1
active
oldest
votes
active
oldest
votes
active
oldest
votes
If you have knowledge that your dependent (y value) is always positive you can use a loss function that constrains your predictions to be in the positive domain as well.
One example supported in XGBoost is Gamma regression (see reg:gamma) or you can design your own loss function like the Mean Squared Log Error, you'd have to derive first and second order derivatives in that case.
answered Nov 19 '18 at 14:17
BarBar
1,3011931
add a comment |
If you have knowledge that your dependent (y value) is always positive you can use a loss function that constrains your predictions to be in the positive domain as well.
One example supported in XGBoost is Gamma regression (see reg:gamma) or you can design your own loss function like the Mean Squared Log Error, you'd have to derive first and second order derivatives in that case.
answered Nov 19 '18 at 14:17
BarBar
1,3011931
add a comment |
If you have knowledge that your dependent (y value) is always positive you can use a loss function that constrains your predictions to be in the positive domain as well.
One example supported in XGBoost is Gamma regression (see reg:gamma) or you can design your own loss function like the Mean Squared Log Error, you'd have to derive first and second order derivatives in that case.
answered Nov 19 '18 at 14:17
BarBar
1,3011931
If you have knowledge that your dependent (y value) is always positive you can use a loss function that constrains your predictions to be in the positive domain as well.
One example supported in XGBoost is Gamma regression (see reg:gamma) or you can design your own loss function like the Mean Squared Log Error, you'd have to derive first and second order derivatives in that case.
answered Nov 19 '18 at 14:17
BarBar
1,3011931
answered Nov 19 '18 at 14:17
BarBar
1,3011931
answered Nov 19 '18 at 14:17
BarBar
1,3011931
answered Nov 19 '18 at 14:17
BarBar
1,3011931
1,3011931
add a comment |
add a comment |
Thanks for contributing an answer to Stack Overflow!
- Please be sure to answer the question. Provide details and share your research!
But avoid …
- Asking for help, clarification, or responding to other answers.
- Making statements based on opinion; back them up with references or personal experience.
To learn more, see our tips on writing great answers.
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
StackExchange.ready(
function ()
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fstackoverflow.com%2fquestions%2f53321395%2fcontrol-that-predictions-are-0-using-gridsearchcv%23new-answer', 'question_page');
);
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
you could use the mean absolute error (MAE)
– makaros
Nov 15 '18 at 23:27
Or you could a use a custom wrapper which just wraps around the
neg_mean_squared_log_errorand clips the negative values to 0 before passing it to the scorer, if that makes sense.– Vivek Kumar
Nov 16 '18 at 4:39