When GBM model saved and loaded doesn't give same predicted values when model loaded again and again
I trained a GBM model (classification) and saved it using:
saveRDS(gbmfit, file='E:/..../gbm_nb.rds')
when I am using above model for scoring data, I load it using:
gbmfit <- readRDS('E:/..../gbm_nb.rds')
and predict:
nb_lapse$PRED <- predict(gbmfit, nb_lapse, type='response', n.trees=2000)
This PRED values should be same each time I run it against same scoring data. But some of these values come different. I have printed these by two runs of above code and the differences are around 3% of records. The difference I have printed as such:
Where ID is the unique number. PRD_ORIG is coming in the first run and PRED_NEW is coming in the second run. In second run I am loading model again. If I am not loading model, instead score data again on already loaded model then I don't see any difference in PRED values.
Have you seen this kind of behavior? I was expecting that this should give same PRED values, for same scoring data, each time I will load model and score it.
Thanks.
r gbm
asked Nov 20 at 17:22
Sanoj
4042515
add a comment |
I trained a GBM model (classification) and saved it using:
saveRDS(gbmfit, file='E:/..../gbm_nb.rds')
when I am using above model for scoring data, I load it using:
gbmfit <- readRDS('E:/..../gbm_nb.rds')
and predict:
nb_lapse$PRED <- predict(gbmfit, nb_lapse, type='response', n.trees=2000)
This PRED values should be same each time I run it against same scoring data. But some of these values come different. I have printed these by two runs of above code and the differences are around 3% of records. The difference I have printed as such:
Where ID is the unique number. PRD_ORIG is coming in the first run and PRED_NEW is coming in the second run. In second run I am loading model again. If I am not loading model, instead score data again on already loaded model then I don't see any difference in PRED values.
Have you seen this kind of behavior? I was expecting that this should give same PRED values, for same scoring data, each time I will load model and score it.
Thanks.
r gbm
asked Nov 20 at 17:22
Sanoj
4042515
add a comment |
I trained a GBM model (classification) and saved it using:
saveRDS(gbmfit, file='E:/..../gbm_nb.rds')
when I am using above model for scoring data, I load it using:
gbmfit <- readRDS('E:/..../gbm_nb.rds')
and predict:
nb_lapse$PRED <- predict(gbmfit, nb_lapse, type='response', n.trees=2000)
This PRED values should be same each time I run it against same scoring data. But some of these values come different. I have printed these by two runs of above code and the differences are around 3% of records. The difference I have printed as such:
Where ID is the unique number. PRD_ORIG is coming in the first run and PRED_NEW is coming in the second run. In second run I am loading model again. If I am not loading model, instead score data again on already loaded model then I don't see any difference in PRED values.
Have you seen this kind of behavior? I was expecting that this should give same PRED values, for same scoring data, each time I will load model and score it.
Thanks.
r gbm
asked Nov 20 at 17:22
Sanoj
4042515
I trained a GBM model (classification) and saved it using:
saveRDS(gbmfit, file='E:/..../gbm_nb.rds')
when I am using above model for scoring data, I load it using:
gbmfit <- readRDS('E:/..../gbm_nb.rds')
and predict:
nb_lapse$PRED <- predict(gbmfit, nb_lapse, type='response', n.trees=2000)
This PRED values should be same each time I run it against same scoring data. But some of these values come different. I have printed these by two runs of above code and the differences are around 3% of records. The difference I have printed as such:
Where ID is the unique number. PRD_ORIG is coming in the first run and PRED_NEW is coming in the second run. In second run I am loading model again. If I am not loading model, instead score data again on already loaded model then I don't see any difference in PRED values.
Have you seen this kind of behavior? I was expecting that this should give same PRED values, for same scoring data, each time I will load model and score it.
Thanks.
r gbm
r gbm
asked Nov 20 at 17:22
Sanoj
4042515
asked Nov 20 at 17:22
Sanoj
4042515
asked Nov 20 at 17:22
Sanoj
4042515
asked Nov 20 at 17:22
Sanoj
4042515
asked Nov 20 at 17:22
Sanoj
4042515
4042515
add a comment |
add a comment |
1 Answer
1
active
oldest
votes
I found the answer. I have one categorical variable, say A, with many categories. In the train data, I have capped those categories based on another variable, say B, in the data frame. In scoring data frame that variable B is not present. Therefore, despite my scoring data goes through capping code, actual capping of categories doesn't happen. R doesn't throw any error if code is not executed. Therefore in scoring data I end up having more categories in categorical variable A than in train data on which model was trained. This was creating problem and I was getting different predicted values for those data points where categories were different, or not present, than in train data. Once I capped scoring data categorical variable as in train data my predicted values are coming same every time.
answered Nov 20 at 21:22
Sanoj
4042515
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%2f53398305%2fwhen-gbm-model-saved-and-loaded-doesnt-give-same-predicted-values-when-model-lo%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
I found the answer. I have one categorical variable, say A, with many categories. In the train data, I have capped those categories based on another variable, say B, in the data frame. In scoring data frame that variable B is not present. Therefore, despite my scoring data goes through capping code, actual capping of categories doesn't happen. R doesn't throw any error if code is not executed. Therefore in scoring data I end up having more categories in categorical variable A than in train data on which model was trained. This was creating problem and I was getting different predicted values for those data points where categories were different, or not present, than in train data. Once I capped scoring data categorical variable as in train data my predicted values are coming same every time.
answered Nov 20 at 21:22
Sanoj
4042515
add a comment |
I found the answer. I have one categorical variable, say A, with many categories. In the train data, I have capped those categories based on another variable, say B, in the data frame. In scoring data frame that variable B is not present. Therefore, despite my scoring data goes through capping code, actual capping of categories doesn't happen. R doesn't throw any error if code is not executed. Therefore in scoring data I end up having more categories in categorical variable A than in train data on which model was trained. This was creating problem and I was getting different predicted values for those data points where categories were different, or not present, than in train data. Once I capped scoring data categorical variable as in train data my predicted values are coming same every time.
answered Nov 20 at 21:22
Sanoj
4042515
add a comment |
I found the answer. I have one categorical variable, say A, with many categories. In the train data, I have capped those categories based on another variable, say B, in the data frame. In scoring data frame that variable B is not present. Therefore, despite my scoring data goes through capping code, actual capping of categories doesn't happen. R doesn't throw any error if code is not executed. Therefore in scoring data I end up having more categories in categorical variable A than in train data on which model was trained. This was creating problem and I was getting different predicted values for those data points where categories were different, or not present, than in train data. Once I capped scoring data categorical variable as in train data my predicted values are coming same every time.
answered Nov 20 at 21:22
Sanoj
4042515
I found the answer. I have one categorical variable, say A, with many categories. In the train data, I have capped those categories based on another variable, say B, in the data frame. In scoring data frame that variable B is not present. Therefore, despite my scoring data goes through capping code, actual capping of categories doesn't happen. R doesn't throw any error if code is not executed. Therefore in scoring data I end up having more categories in categorical variable A than in train data on which model was trained. This was creating problem and I was getting different predicted values for those data points where categories were different, or not present, than in train data. Once I capped scoring data categorical variable as in train data my predicted values are coming same every time.
answered Nov 20 at 21:22
Sanoj
4042515
answered Nov 20 at 21:22
Sanoj
4042515
answered Nov 20 at 21:22
Sanoj
4042515
answered Nov 20 at 21:22
Sanoj
4042515
4042515
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.
Some of your past answers have not been well-received, and you're in danger of being blocked from answering.
Please pay close attention to the following guidance:
- 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%2f53398305%2fwhen-gbm-model-saved-and-loaded-doesnt-give-same-predicted-values-when-model-lo%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
