Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                
Scribd
Upload
146 views

1.4 Intro To Need of Estimation and Validation PDF

The document discusses several key points regarding data science model evaluation and validation: 1) It explains the importance of selecting optimal hyperparameters and evaluating model performance on unseen data to estimate generalization error. 2) Learning curves are discussed as a useful tool for understanding bias and variance in models and how more data can help address overfitting. 3) When data is limited, strategies like stratifying samples or data augmentation are recommended over separating into separate training, validation, and test sets. K-fold cross validation is also introduced.

Uploaded by

Dhairya Thakkar
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
146 views

1.4 Intro To Need of Estimation and Validation PDF

The document discusses several key points regarding data science model evaluation and validation: 1) It explains the importance of selecting optimal hyperparameters and evaluating model performance on unseen data to estimate generalization error. 2) Learning curves are discussed as a useful tool for understanding bias and variance in models and how more data can help address overfitting. 3) When data is limited, strategies like stratifying samples or data augmentation are recommended over separating into separate training, validation, and test sets. K-fold cross validation is also introduced.

Uploaded by

Dhairya Thakkar
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 18

Introduction To Need Of Estimation

And Validation For Added Value Due


To Data Science
Vaibhav P. Vasani

Assistant Professor
Department of Computer Engineering
K. J. Somaiya College of Engineering
Somaiya Vidyavihar University

2/17/2022 vaibhav.vasani@gmail.com 1
Introduction To Need Of Estimation
And Validation For Added Value Due
To Data Science

2/17/2022 2
Model selection at different scales

• Hyperparameters are the parameters of the learning


method itself which have to specify a priori, i.e., before
model fitting.
• In contrast, model parameters are parameters which arise
as a result of the fit.

• Example
o In a logistic regression model, for example, the regularization
strength (as well as the regularization type, if any) is a
hyperparameter which has to be specified prior to the fitting,
while the coefficients of the fitted model are model parameters.
Finding the right hyperparameters for a model can be crucial for
the model performance on given data.

2/17/2022 3
• the best learning method (and their corresponding
“optimal” hyperparameters) from a set of eligible
machine learning methods.
o algorithm selection.

2/17/2022 4
Model evaluation

• Model evaluation aims at estimating the


generalization error of the selected model, i.e., how
well the selected model performs on unseen data.

2/17/2022 5
Fitting of curve slide

2/17/2022 6
If data is not an issue

• The recommended strategy for model selection


depends on the amount of data available. If plenty of
data is available, It can split the data into several
parts, each serving a special purpose.
• For instance, for hyperparameter tuning we may split
the data into three sets: train / validation / test.
• There is no general rule as to how the data should be
split. A typical split is e.g. 50%/25%/25%.

2/17/2022 8
Learning curves, and why they are
useful
• In a learning curve, the performance of a model both
on the training and validation set is plotted as a
function of the training set size.

2/17/2022 9
• The training score (performance on the training set) decreases with
increasing training set size while the validation score increases at the same
time.
• High training score and low validation score at the same time indicates that
the model has overfit the data, i.e., has adapted too well to the specific
training set samples.
• As the training set increases, overfitting decreases, and the validation score
increases.
• Especially for data-hungry machine learning models, the learning curve
might not yet have reached a plateau at the given training set size, which
means the generalization error might still decrease when providing more
data to the model.
• Hence, it seems reasonable to increase the training set (by adding the
validation set) before estimating the generalization error on the test set, and
to further take advantage of the test set data for model fitting before
shipping the model. Whether or not this strategy is needed depends strongly
on the slope of the learning curve at the initial training set size.

2/17/2022 10
• Learning curves further allow to easily illustrate the concept of
(statistical) bias and variance. Bias in this context refers to
erroneous (e.g. simplifying) model assumptions, which can cause the
model to underfit the data. A high-bias model does not adequately
capture the structure present in the data. Variance on the other hand
quantifies how much the model varies as we change the training
data. A high-variance model is very sensitive to small fluctuations in
the training data, which can cause the model to overfit. The amount
of bias and variance can be estimated using learning curves: A model
exhibits high variance, but low bias if the training score plateaus at a
high level while the validation score at a low level, i.e., if there is a
large gap between training and validation score. A model with low
variance but high bias, in contrast, is a model where both training
and validation score are low, but similar. Very simple models are
high-bias, low-variance while with increasing model complexity they
become low-bias, high-variance.

2/17/2022 11
2/17/2022 12
• Bias
• Bias is the error resulting from the difference between the
expected value(s) of a model and the actual (or “correct”)
value(s) for which we want to predict over multiple
iterations. In the scientific concepts of accuracy and
precision, bias is very similar to accuracy.
• Variance
• Variance is defined as the error resulting from the
variability between different data predictions in a model.
In variance, the correct value(s) don’t matter as much as
the range of differences in value between the predictions.
Variance also comes into play more when we run multiple
model creation trials.

2/17/2022 13
2/17/2022 14
Divide and conquer — but do it
carefully
• Training, validation, and test set are sampled from
the same distribution.
• Ensure before model building that the distribution of
the data is not affected by partitioning your data.
o Example

• use stratified sampling

2/17/2022 15
If all you have is small data

• split the data into two sets, a training and a test set.
• Augmentation of Data

2/17/2022 17
K-Fold Cross Validation

• shuffles the data and splits it into k number of folds


(groups).

2/17/2022 18
2/17/2022 19
2/17/2022 20

You might also like