Dask categorical encoding applied to train and test dataframes
$begingroup$
I am relatively new to dask and I’m trying to ensure my training and test dataframes share the same categorical information for label and onehot encoding. Obviously if this is not the case then my test predictions will likely be completely incorrect.
Approach 1 (failed)
I concatenated the training and test sets together with
train_test = dd.concat([train,test], axis=0)
I then completed the label and one hot encoding successfully on the new combined dataframe (test_train).
train_test = train_test.categorize(columns=categorical_list+onehot_list)
oe = OrdinalEncoder()
oe.fit(train_test[categorical_list])
train_test[categorical_list] = oe.transform(train_test[categorical_list])
enc = OneHotEncoder()
enc.fit(train_test[onehot_list])
onehot_df = enc.transform(train_test[onehot_list])
train_test[onehot_cols]=onehot_df
However, I then found that I couldn’t then split the dataframe back into a training and test set with: -
train = train_test[:len_train]
test = train_test[len_train:]
len(train) #should be 8921483, ends up being higher (train+test)
This code runs but it fails silently without an error. The length of the dataframe remained unchanged. I realise row-wise operations are hard with dask but without being able to slice the two sets back to train and test I’m at a brick-wall with this approach.
Note: I cannot compute() the entire training dataframe as the resulting pandas dataframe will not fit in memory (it’s a big dataset). I plan to fit my model on a random sample of the train data.
Approach 2 (failed)
I kept train and test as separate dataframes. I tried to set the test categories to be the same as the training categories, but the syntax that might work in Panda’s does not behave the same in dask.
for n in test.columns:
if (n in train.columns) and (train[n].dtype.name=='category'):
test[n] = test[n].astype('category').cat.as_ordered()
test[n].cat.set_categories(train[n].cat.categories, ordered=True, inplace=True)
It’s possible that both of my approaches are at fault and I’m missing something basic. I just wish to ensure that my categorical label and onehot encoding is the same in the test set as the training set. I’d be very grateful for any help!
machine-learning bigdata categorical-data
New contributor
$endgroup$
add a comment |
$begingroup$
I am relatively new to dask and I’m trying to ensure my training and test dataframes share the same categorical information for label and onehot encoding. Obviously if this is not the case then my test predictions will likely be completely incorrect.
Approach 1 (failed)
I concatenated the training and test sets together with
train_test = dd.concat([train,test], axis=0)
I then completed the label and one hot encoding successfully on the new combined dataframe (test_train).
train_test = train_test.categorize(columns=categorical_list+onehot_list)
oe = OrdinalEncoder()
oe.fit(train_test[categorical_list])
train_test[categorical_list] = oe.transform(train_test[categorical_list])
enc = OneHotEncoder()
enc.fit(train_test[onehot_list])
onehot_df = enc.transform(train_test[onehot_list])
train_test[onehot_cols]=onehot_df
However, I then found that I couldn’t then split the dataframe back into a training and test set with: -
train = train_test[:len_train]
test = train_test[len_train:]
len(train) #should be 8921483, ends up being higher (train+test)
This code runs but it fails silently without an error. The length of the dataframe remained unchanged. I realise row-wise operations are hard with dask but without being able to slice the two sets back to train and test I’m at a brick-wall with this approach.
Note: I cannot compute() the entire training dataframe as the resulting pandas dataframe will not fit in memory (it’s a big dataset). I plan to fit my model on a random sample of the train data.
Approach 2 (failed)
I kept train and test as separate dataframes. I tried to set the test categories to be the same as the training categories, but the syntax that might work in Panda’s does not behave the same in dask.
for n in test.columns:
if (n in train.columns) and (train[n].dtype.name=='category'):
test[n] = test[n].astype('category').cat.as_ordered()
test[n].cat.set_categories(train[n].cat.categories, ordered=True, inplace=True)
It’s possible that both of my approaches are at fault and I’m missing something basic. I just wish to ensure that my categorical label and onehot encoding is the same in the test set as the training set. I’d be very grateful for any help!
machine-learning bigdata categorical-data
New contributor
$endgroup$
add a comment |
$begingroup$
I am relatively new to dask and I’m trying to ensure my training and test dataframes share the same categorical information for label and onehot encoding. Obviously if this is not the case then my test predictions will likely be completely incorrect.
Approach 1 (failed)
I concatenated the training and test sets together with
train_test = dd.concat([train,test], axis=0)
I then completed the label and one hot encoding successfully on the new combined dataframe (test_train).
train_test = train_test.categorize(columns=categorical_list+onehot_list)
oe = OrdinalEncoder()
oe.fit(train_test[categorical_list])
train_test[categorical_list] = oe.transform(train_test[categorical_list])
enc = OneHotEncoder()
enc.fit(train_test[onehot_list])
onehot_df = enc.transform(train_test[onehot_list])
train_test[onehot_cols]=onehot_df
However, I then found that I couldn’t then split the dataframe back into a training and test set with: -
train = train_test[:len_train]
test = train_test[len_train:]
len(train) #should be 8921483, ends up being higher (train+test)
This code runs but it fails silently without an error. The length of the dataframe remained unchanged. I realise row-wise operations are hard with dask but without being able to slice the two sets back to train and test I’m at a brick-wall with this approach.
Note: I cannot compute() the entire training dataframe as the resulting pandas dataframe will not fit in memory (it’s a big dataset). I plan to fit my model on a random sample of the train data.
Approach 2 (failed)
I kept train and test as separate dataframes. I tried to set the test categories to be the same as the training categories, but the syntax that might work in Panda’s does not behave the same in dask.
for n in test.columns:
if (n in train.columns) and (train[n].dtype.name=='category'):
test[n] = test[n].astype('category').cat.as_ordered()
test[n].cat.set_categories(train[n].cat.categories, ordered=True, inplace=True)
It’s possible that both of my approaches are at fault and I’m missing something basic. I just wish to ensure that my categorical label and onehot encoding is the same in the test set as the training set. I’d be very grateful for any help!
machine-learning bigdata categorical-data
New contributor
$endgroup$
I am relatively new to dask and I’m trying to ensure my training and test dataframes share the same categorical information for label and onehot encoding. Obviously if this is not the case then my test predictions will likely be completely incorrect.
Approach 1 (failed)
I concatenated the training and test sets together with
train_test = dd.concat([train,test], axis=0)
I then completed the label and one hot encoding successfully on the new combined dataframe (test_train).
train_test = train_test.categorize(columns=categorical_list+onehot_list)
oe = OrdinalEncoder()
oe.fit(train_test[categorical_list])
train_test[categorical_list] = oe.transform(train_test[categorical_list])
enc = OneHotEncoder()
enc.fit(train_test[onehot_list])
onehot_df = enc.transform(train_test[onehot_list])
train_test[onehot_cols]=onehot_df
However, I then found that I couldn’t then split the dataframe back into a training and test set with: -
train = train_test[:len_train]
test = train_test[len_train:]
len(train) #should be 8921483, ends up being higher (train+test)
This code runs but it fails silently without an error. The length of the dataframe remained unchanged. I realise row-wise operations are hard with dask but without being able to slice the two sets back to train and test I’m at a brick-wall with this approach.
Note: I cannot compute() the entire training dataframe as the resulting pandas dataframe will not fit in memory (it’s a big dataset). I plan to fit my model on a random sample of the train data.
Approach 2 (failed)
I kept train and test as separate dataframes. I tried to set the test categories to be the same as the training categories, but the syntax that might work in Panda’s does not behave the same in dask.
for n in test.columns:
if (n in train.columns) and (train[n].dtype.name=='category'):
test[n] = test[n].astype('category').cat.as_ordered()
test[n].cat.set_categories(train[n].cat.categories, ordered=True, inplace=True)
It’s possible that both of my approaches are at fault and I’m missing something basic. I just wish to ensure that my categorical label and onehot encoding is the same in the test set as the training set. I’d be very grateful for any help!
machine-learning bigdata categorical-data
machine-learning bigdata categorical-data
New contributor
New contributor
New contributor
asked 53 mins ago
Nigel AdamsNigel Adams
1
1
New contributor
New contributor
add a comment |
add a comment |
0
active
oldest
votes
Your Answer
StackExchange.ifUsing("editor", function () {
return StackExchange.using("mathjaxEditing", function () {
StackExchange.MarkdownEditor.creationCallbacks.add(function (editor, postfix) {
StackExchange.mathjaxEditing.prepareWmdForMathJax(editor, postfix, [["$", "$"], ["\\(","\\)"]]);
});
});
}, "mathjax-editing");
StackExchange.ready(function() {
var channelOptions = {
tags: "".split(" "),
id: "557"
};
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: false,
noModals: true,
showLowRepImageUploadWarning: true,
reputationToPostImages: null,
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
});
}
});
Nigel Adams is a new contributor. Be nice, and check out our Code of Conduct.
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%2fdatascience.stackexchange.com%2fquestions%2f44314%2fdask-categorical-encoding-applied-to-train-and-test-dataframes%23new-answer', 'question_page');
}
);
Post as a guest
Required, but never shown
0
active
oldest
votes
0
active
oldest
votes
active
oldest
votes
active
oldest
votes
Nigel Adams is a new contributor. Be nice, and check out our Code of Conduct.
Nigel Adams is a new contributor. Be nice, and check out our Code of Conduct.
Nigel Adams is a new contributor. Be nice, and check out our Code of Conduct.
Nigel Adams is a new contributor. Be nice, and check out our Code of Conduct.
Thanks for contributing an answer to Data Science Stack Exchange!
- 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.
Use MathJax to format equations. MathJax reference.
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%2fdatascience.stackexchange.com%2fquestions%2f44314%2fdask-categorical-encoding-applied-to-train-and-test-dataframes%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