We Keep Coding

How is it that the accuracy score for 10-fold cross validation is worst than for a 90-10 train_test_split using sklearn?

The task is binary classification via a neural network. The data is present in a dictionary, that contains composite names (as the key) of each entries and the labels (0 or 1, as the third element in the vector value). The first and second elements are the two parts of the composite name, which are used later to extract the corresponding features.
In both cases, the dictionary is transformed into two arrays for the purpose of performing a balanced undersampling of the majority class (that is present in 66% of the data):
data_for_sampling = np.asarray([key for key in list(data.keys())])
labels_for_sampling = [element[2] for element in list(data.values())]
sampler = RandomUnderSampler(sampling_strategy = 'majority')
data_sampled, label_sampled = sampler.fit_resample(data_for_sampling.reshape(-1, 1), labels_for_sampling)
Then the resampled arrays of names and labels are used to create train and test sets via the Kfold method:
kfolder = KFold(n_splits = 10, shuffle = True)
kfolder.get_n_splits(data_sampled)
for train_index, test_index in kfolder.split(data_sampled):
data_train, data_test = data_sampled[train_index], data_sampled[test_index]
Or the train_test_split method:
data_train, data_test, label_train, label_test = train_test_split(data_sampled, label_sampled, test_size = 0.1, shuffle = True)
Finally, the names from data_train and data_test are used to re-extract the relevant entries (by key) from the original dictionary, that is then used to gather the features of those entries. As far as I'm concerned, a single split of the 10-fold sets should provide similar train-test data distribution as the 90-10 train_test_split, and that seems to be true during training, where both training sets result in ~0.82 accuracy after only one epoch, run separately with model.fit(). However, when the corresponding models are evaluated using model.evaluate() on the test sets after said epoch, the set from train_test_split gives a score of ~0.86, while the set from Kfold is ~0.72. I have done numerous testing to see if it's just a bad random seed, which is not bounded, but the results stayed the same. The sets also have correctly balanced label distributions and seemingly well-shuffled entries.

As it turns out, the problem originates from a combination of sources:
While shuffle = True in the train_test_split() method properly shuffles the provided data first, then splits it into the desired parts, the shuffle = True in the Kfold method only results in the randomly built folds, however the data within the folds remains ordered.
This is something the documentation points out, thanks to this post:
https://github.com/scikit-learn/scikit-learn/issues/16068
Now, during learning, my custom train function applies shuffle again on the train data, just to be sure, however it does not shuffle the test data. Moreover, model.evaluate() defaults to batch_size = 32, if no parameter is given, which paired with the ordered test data resulted in the discrepancy in the validation accuracy. The test data is indeed flawed in the sense that it contains large portion of "hard-to-predict" entries, which were clustered together thanks to the ordering and seems like they dragged down the average accuracy in the results. Given a completed run across all N folds, as pointed out by TC Arlen, may have indeed given a more precise estimation in the end, but I've expected closer results after only one fold, which lead to the discovery of this problem.

Depending on the amount of noise in the data and on the size of the dataset, this could be expected behavior to see scores on out of sample data to deviate by this amount. One split is not guaranteed to be just like any other split, which is why you have 10 in the first place and then average across all results.
What you should trust to be the most generalizable is not any one given split (whether that comes from one of the 10 folds or train_test_split()), but what is far more trustworthy is the average result across all N folds.
Digging deeper into the data could reveal whether there is some reason why one or more splits deviate so much from another. For example, perhaps there is some feature in your data (e.g. "date the sample was collected" and the collection methodology changed from month to month) that makes the data differ from one another in a biased way. If that is the case, you should use a stratified test split (in your CV as well) (see the scikit-learn documentation on that) so you can get a more unbiased grouping of your data.

Pairwise comparisons for model training/testing - how to parameter tune?

For some reasons, I have base dataframes of the following structure
print(df1.shape)
display(df1.head())
print(df2.shape)
display(df2.head())
Where the top dataframe is my features set and my bottom is the output set. To turn this into a problem that is amenable to data modeling I first do:
x_train, x_test, y_train, y_test = train_test_split(df1, df2, train_size = 0.8)
I then have a split for 80% training and 20% testing.
Since the output set (df2; y_test/y_train) is individual measurements with no inherent meaning on their own, I calculate pairwise distances between the labels to generate a single output value denoting the pairwise distances between observations using (the distances are computed after z-scoring; the z-scoring code isn't described here but it is done):
y_train = pdist(y_train, 'euclidean')
y_test = pdist(y_test, 'euclidean')
Similarly I then apply this strategy to the features set to generate pairwise distances between individual observations of each of the instances of each feature.
def feature_distances(input_vector):
modified_vector = np.array(input_vector).reshape(-1,1)
vector_distances = pdist(modified_vector, 'euclidean')
vector_distances = pd.Series(vector_distances)
return vector_distances
x_train = x_train.apply(feature_distances, axis = 0)
x_test = x_test.apply(feature_distances, axis = 0)
I then proceed to train & test all of my models.
For now I am trying linear regression , random forest, xgboost.
Is there any easy way to implement a cross validation scheme in my dataset?
Since my problem requires calculating pairwise distances between observations, I am struggling to identify an easy way to do cross validation schemes to optimize parameter tuning.
GridsearchCV doesn't quite work here since in each instance of the test/train split, distances have to be recomputed to avoid contamination of test with train.
Hope it's clear!

First, what I understood from the shape of your data frames that you have 42 samples and 1643 features in the input, and each output vector consists of 392 values.
Huge Input: In case, you are sure that your problem has 1643 features, you might need to use PCA to reduce the dimensionality instead of pairwise distance. You should collect more samples instead of 42 samples to avoid overfitting because it is not enough data to train and test your model.
Huge Output: you could use sampled_softmax_loss to speed up the training process as mentioned in TensorFlow documentation . You could also read this here. In case, you do not want to follow this approach, you can continue training with this output but it takes some time.

x_train, x_test, y_train, y_test = train_test_split(X, y, test_size=0.8, random_state=n)
here X is independent feature, y is dependent feature means what you actually want to predict - it could be label or continuous value. We used train_test_split on train dataset and we are using (x_train, y_train) to train model and (x_test, y_test) to test model to ensure performance of model on unknown data(x_test, y_test). In your case you have given y as df2 which is wrong just figure out your target feature and give it as y and there is no need to split test data.

Building ML classifier with imbalanced data

I have a dataset with 1400 obs and 19 columns. The Target variable has values 1 (value that I am most interested in) and 0. The distribution of classes shows imbalance (70:30).
Using the code below I am getting weird values (all 1s). I am not figuring out if this is due to a problem of overfitting/imbalance data or to feature selection (I used Pearson correlation since all values are numeric/boolean).
I am thinking that the steps followed are wrong.
import numpy as np
import math
import sklearn.metrics as metrics
from sklearn.metrics import f1_score
y = df['Label']
X = df.drop('Label',axis=1)
def create_cv(X,y):
if type(X)!=np.ndarray:
X=X.values
y=y.values
test_size=1/5
proportion_of_true=y[y==1].shape[0]/y.shape[0]
num_test_samples=math.ceil(y.shape[0]*test_size)
num_test_true_labels=math.floor(num_test_samples*proportion_of_true)
num_test_false_labels=math.floor(num_test_samples-num_test_true_labels)
y_test=np.concatenate([y[y==0][:num_test_false_labels],y[y==1][:num_test_true_labels]])
y_train=np.concatenate([y[y==0][num_test_false_labels:],y[y==1][num_test_true_labels:]])
X_test=np.concatenate([X[y==0][:num_test_false_labels] ,X[y==1][:num_test_true_labels]],axis=0)
X_train=np.concatenate([X[y==0][num_test_false_labels:],X[y==1][num_test_true_labels:]],axis=0)
return X_train,X_test,y_train,y_test
X_train,X_test,y_train,y_test=create_cv(X,y)
X_train,X_crossv,y_train,y_crossv=create_cv(X_train,y_train)
tree = DecisionTreeClassifier(max_depth = 5)
tree.fit(X_train, y_train)
y_predict_test = tree.predict(X_test)
print(classification_report(y_test, y_predict_test))
f1_score(y_test, y_predict_test)
Output:
precision recall f1-score support
0 1.00 1.00 1.00 24
1 1.00 1.00 1.00 70
accuracy 1.00 94
macro avg 1.00 1.00 1.00 94
weighted avg 1.00 1.00 1.00 94
Has anyone experienced similar issues in building a classifier when data has imbalance, using CV and/or under sampling? Happy to share the whole dataset, in case you might want to replicate the output.
What I would like to ask you for some clear answer to follow that can show me the steps and what I am doing wrong.
I know that, to reduce overfitting and work with balance data, there are some methods such as random sampling (over/under), SMOTE, CV. My idea is
Split the data on train/test taking into account imbalance
Perform CV on trains set
Apply undersampling only on a test fold
After the model has been chosen with the help of CV, undersample the train set and train the classifier
Estimate the performance on the untouched test set
(f1-score)
as also outlined in this question: CV and under sampling on a test fold .
I think the steps above should make sense, but happy to receive any feedback that you might have on this.

When you have imbalanced data you have to perform stratification. The usual way is to oversample the class that has less values.
Another option is to train your algorithm with less data. If you have a good dataset that should not be a problem. In this case you grab first the samples from the less represented class use the size of the set to compute how many samples to get from the other class:
This code may help you split your dataset that way:
def split_dataset(dataset: pd.DataFrame, train_share=0.8):
"""Splits the dataset into training and test sets"""
all_idx = range(len(dataset))
train_count = int(len(all_idx) * train_share)
train_idx = random.sample(all_idx, train_count)
test_idx = list(set(all_idx).difference(set(train_idx)))
train = dataset.iloc[train_idx]
test = dataset.iloc[test_idx]
return train, test
def split_dataset_stratified(dataset, target_attr, positive_class, train_share=0.8):
"""Splits the dataset as in `split_dataset` but with stratification"""
data_pos = dataset[dataset[target_attr] == positive_class]
data_neg = dataset[dataset[target_attr] != positive_class]
if len(data_pos) < len(data_neg):
train_pos, test_pos = split_dataset(data_pos, train_share)
train_neg, test_neg = split_dataset(data_neg, len(train_pos)/len(data_neg))
# set.difference makes the test set larger
test_neg = test_neg.iloc[0:len(test_pos)]
else:
train_neg, test_neg = split_dataset(data_neg, train_share)
train_pos, test_pos = split_dataset(data_pos, len(train_neg)/len(data_pos))
# set.difference makes the test set larger
test_pos = test_pos.iloc[0:len(test_neg)]
return train_pos.append(train_neg).sample(frac = 1).reset_index(drop = True), \
test_pos.append(test_neg).sample(frac = 1).reset_index(drop = True)
Usage:
train_ds, test_ds = split_dataset_stratified(data, target_attr, positive_class)
You can now perform cross validation on train_ds and evaluate your model in test_ds.

There is another solution that is in the model-level - using models that support weights of samples, such as Gradient Boosted Trees. Of those, CatBoost is usually the best as its training method leads to less leakage (as described in their article).
Example code:
from catboost import CatBoostClassifier
y = df['Label']
X = df.drop('Label',axis=1)
label_ratio = (y==1).sum() / (y==0).sum()
model = CatBoostClassifier(scale_pos_weight = label_ratio)
model.fit(X, y)
And so forth.
This works because Catboost treats each sample with a weight, so you can determine class weights in advance (scale_pos_weight).
This is better than downsampling, and is technically equal to oversampling (but requires less memory).
Also, a major part of treating imbalanced data, is making sure your metrics are weighted as well, or at least well-defined, as you might want equal performance (or skewed performance) on these metrics.
And if you want a more visual output than sklearn's classification_report, you can use one of the Deepchecks built-in checks (disclosure - I'm one of the maintainers):
from deepchecks.checks import PerformanceReport
from deepchecks import Dataset
PerformanceReport().run(Dataset(train_df, label='Label'), Dataset(test_df, label='Label'), model)

your implementation of stratified train/test creation is not optimal, as it lacks randomness. Very often data comes in batches, so it is not a good practice to take sequences of data as is, without shuffling.
as #sturgemeister mentioned, classes ratio 3:7 is not critical, so you should not worry too much of class imbalance. When you artificially change data balance in training you will need to compensate it by multiplication by prior for some algorithms.
as for your "perfect" results either your model overtrained or the model is indeed classifies the data perfectly. Use different train/test split to check this.
another point: your test set is only 94 data points. It is definitely not 1/5 of 1400. Check your numbers.
to get realistic estimates, you need lots of test data. This is the reason why you need to apply Cross Validation strategy.
as for general strategy for 5-fold CV I suggest following:
split your data to 5 folds with respect to labels (this is called stratified split and you can use StratifiedShuffleSplit function)
take 4 splits and train your model. If you want to use under/oversampling, modify the data in those 4 training splits.
apply the model to the remaining part. Do not under/over sample data in the test part. This way you get realistic performance estimate. Save the results.
repeat 2. and 3. for all test splits (totally 5 times obviously). Important: do not change parameters (e.g. tree depth) of the model when training - they should be the same for all splits.
now you have all your data points tested without being trained on them. This is the core idea of cross validation. Concatenate all the saved results, and estimate the performance .

Cross-validation or held-out set
First of all, you are not doing cross-validation. You are splitting your data in a train/validation/test set, which is good, and often sufficient when the number of training samples is large (say, >2e4). However, when the number of samples is small, which is your case, cross-validation becomes useful.
It is explained in depth in scikit-learn's documentation. You will start by taking out a test set from your data, as your create_cv function does. Then, you split the rest of the training data in e.g. 3 splits. Then, you do, for i in {1, 2, 3}: train on data j != i, evaluate on data i. The documentation explains it with prettier and colorful figures, you should have a look! It can be quite cumbersome to implement, but hopefully scikit does it out of the box.
As for the dataset being unbalanced, it is a very good idea to keep the same ratio of labels in each set. But again, you can let scikit handle it for you!
Purpose
Also, the purpose of cross-validation is to choose the right values for the hyper-parameters. You want the right amount of regularization, not too big (under-fitting) nor too small (over-fitting). If you're using a decision tree, the maximum depth (or the minimum number of samples per leaf) is the right metric to consider to estimate the regularization of your method.
Conclusion
Simply use GridSearchCV. You will have cross-validation and label balance done for you.
from sklearn.model_selection import train_test_split
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=1/5, stratified=True)
tree = DecisionTreeClassifier()
parameters = {'min_samples_leaf': [1, 5, 10]}
clf = GridSearchCV(svc, parameters, cv=5) # Specifying cv does StratifiedShuffleSplit, see documentation
clf.fit(iris.data, iris.target)
sorted(clf.cv_results_.keys())
You can also replace the cv variable by a fancier shuffler, such as StratifiedGroupKFold (no intersection between groups).
I would also advise looking towards random trees, which are less interpretable but said to have better performances in practice.

Just wanted to add thresholding and cost sensitive learning to the list of possible approaches mentioned by the others. The former is well described here and consists in finding a new threshold for classifying positive vs negative classes (generally is 0.5 but it can be treated as an hyper parameter). The latter consists on weighting the classes to cope with their unbalancedness. This article was really useful to me to understand how to deal with unbalanced data sets. In it, you can find also cost sensitive learning with a specific explanation using decision tree as a model. Also all other approaches are really nicely reviewed including: Adaptive Synthetic Sampling, informed undersampling etc.

Logistic Regression - Python?

Could you briefly describe me what the below lines of code mean. This is the code of logistic regression in Python.
What means size =0.25 and random_state = 0 ? And what is train_test_split ? What was done in this line of code ?
X_train,X_test,y_train,y_test = train_test_split(X,y,test_size=0.25,random_state=0)
And what was done in these lines of code ?
logistic_regression= LogisticRegression()
logistic_regression.fit(X_train,y_train)
y_pred=logistic_regression.predict(X_test)

Have a look at the description of the function here:
random_state sets the seed for the random number generator to give you the same result with each run, especially useful in education settings to give everyone an identical result.
test_size refers to the proportion used in the test split, here 75% of the data is used for training, 25% is used for testing the model.
The other lines simply run the logistic regression on the training dataset. You then use the test dataset to check the goodness of the fitted regression.

What means size =0.25 and random_state = 0 ?
test_size=0.25 -> 25% split of training and test data.
random_state = 0 -> for reproducible results this can be any number.
What was done in this line of code ?
Splits X and y into X_train, X_test, y_train, y_test
And what was done in these lines of code ?
Trains the logistic regression model through the fit(X_train, y_train) and then makes predictions on the test set X_test.
Later you probably compare y_pred to y_test to see what the accuracy of the model is.

Based on the documentation:
test_size : float, int or None, optional (default=None)
If float, should be between 0.0 and 1.0 and represent the proportion of the dataset to include in the test split. If int, represents the absolute number of test samples. If None, the value is set to the complement of the train size. If train_size is also None, it will be set to 0.25.
This gives you the split between your train data and test data, if you have in total 1000 data points, a test_size=0.25 would mean that you have:
750 data points for train
250 data points for test
The perfect size is still under discussions, for large datasets (1.000.000+ ) I currently prefer to set it to 0.1. And even before I have another validation dataset, which I will keep completly out until I decided to run the algorithm.
random_state : int, RandomState instance or None, optional
(default=None)
If int, random_state is the seed used by the random number generator; If RandomState instance, random_state is the random number generator; If None, the random number generator is the RandomState instance used by np.random.
For machine learning you should set this to a value, if you set it, you will have the chance to open your programm on another day and still produce the same results, normally random_state is also in all classifiers/regression models avaiable, so that you can start working and tuning, and have it reproducible,
To comment your regression:
logistic_regression= LogisticRegression()
logistic_regression.fit(X_train,y_train)
y_pred=logistic_regression.predict(X_test)
Will load your Regression, for python this is only to name it
Will fit your logistic regression based on your training set, in this example it will use 750 datsets to train the regression. Training means, that the weights of logistic regression will be minimized with the 750 entries, that the estimat for your y_train fits
This will use the learned weights of step 2 to do an estimation for y_pred with the X_test
After that you can test your results, you now have a y_pred which you calculated and the real y_test, you can know calculate some accuracy scores and the how good the regression was trained.

This line line:
X_train,X_test,y_train,y_test = train_test_split(X,y,test_size=0.25,random_state=0)
divides your source into train and test set, 0.25 shows 25% of the source will be used for test and remaining will be used for training.
For, random_state = 0, here is a brief discussion.
A part from above link:
if you use random_state=some_number, then you can guarantee that the
output of Run 1 will be equal to the output of Run 2,
logistic_regression= LogisticRegression() #Creates logistic regressor
Calculates some values for your source. Recommended read
logistic_regression.fit(X_train,y_train)
A part from above link:
Here the fit method, when applied to the training dataset,learns the
model parameters (for example, mean and standard deviation)
....
It doesn't matter what the actual random_state number is 42, 0, 21, ... The important thing is that everytime you use 42, you will always get the same output the first time you make the split. This is useful if you want reproducible results,
Perform prediction on test set based on the learning from training set.
y_pred=logistic_regression.predict(X_test)

X_train,X_test,y_train,y_test = train_test_split(X,y,test_size=0.25,random_state=0)
Above line splits your data into training and testing data randomly
X is your dataset minus output variable
y is your output variable
test_size=0.25 means you are dividing data into 75%-25% where 25% is your testing dataset
random_state is used for generating same sample again when you run the code
Refer train-test-split documentation

Why should we perform a Kfold cross validation on test set??

I was working on a knearest neighbours problem set. I couldn't understand why are they performing K fold cross validation on test set?? Cant we directly test how well our best parameter K performed on the entire test data? rather than doing a cross validation?
iris = sklearn.datasets.load_iris()
X = iris.data
Y = iris.target
X_train, X_test, Y_train, Y_test = sklearn.cross_validation.train_test_split(
X, Y, test_size=0.33, random_state=42)
k = np.arange(20)+1
parameters = {'n_neighbors': k}
knn = sklearn.neighbors.KNeighborsClassifier()
clf = sklearn.grid_search.GridSearchCV(knn, parameters, cv=10)
clf.fit(X_train, Y_train)
def computeTestScores(test_x, test_y, clf, cv):
kFolds = sklearn.cross_validation.KFold(test_x.shape[0], n_folds=cv)
scores = []
for _, test_index in kFolds:
test_data = test_x[test_index]
test_labels = test_y[test_index]
scores.append(sklearn.metrics.accuracy_score(test_labels, clf.predict(test_data)))
return scores
scores = computeTestScores(test_x = X_test, test_y = Y_test, clf=clf, cv=5)

TL;DR
Did you ever have a science teacher who said, 'any measurement without error bounds is meaningless?'
You might worry that the score on using your fitted, hyperparameter optimized, estimator on your test set is a fluke. By doing a number of tests on a randomly chosen subsample of the test set you get a range of scores; you can report their mean and standard deviation etc. This is, hopefully, a better proxy for how the estimator will perform on new data from the wild.
The following conceptual model may not apply to all estimators but it is a useful to bear in mind. You end up needing 3 subsets of your data. You can skip to the final paragraph if the numbered points are things you are already happy with.
Training your estimator will fit some internal parameters that you need not ever see directly. You optimize these by training on the training set.
Most estimators also have hyperparameters (number of neighbours, alpha for Ridge, ...). Hyperparameters also need to be optimized. You need to fit them to a different subset of your data; call it the validation set.
Finally, when you are happy with the fit of both the estimator's internal parameters and the hyperparmeters, you want to see how well the fitted estimator predicts on new data. You need a final subset (the test set) of your data to figure out how well the training and hyperparameter optimization went.
In lots of cases the partitioning your data into 3 means you don't have enough samples in each subset. One way around this is to randomly split the training set a number of times, fit hyperparameters and aggregate the results. This also helps stop your hyperparameters being over-fit to a particular validation set. K-fold cross-validation is one strategy.
Another use for this splitting a data set at random is to get a range of results for how your final estimator did. By splitting the test set and computing the score you get a range of answers to 'how might we do on new data'. The hope is that this is more representative of what you might see as real-world novel data performance. You can also get a standard deviation for you final score. This appears to be what the Harvard cs109 gist is doing.

If you make a program that adapts to input, then it will be optimal for the input you adapted it to.
This leads to a problem known as overfitting.
In order to see if you have made a good or a bad model, you need to test it on some other data that is not what you used to make the model. This is why you separate your data into 2 parts.