stochasticLogisticRegression

Introduced in: v20.1

This function implements stochastic logistic regression. It can be used for binary classification problem, supports the same custom parameters as stochasticLinearRegression and works the same way.

Usage

The function is used in two steps:

Fitting

For fitting a query like this can be used:

CREATE TABLE IF NOT EXISTS train_data
(
    param1 Float64,
    param2 Float64,
    target Float64
) ENGINE = Memory;

CREATE TABLE your_model ENGINE = Memory AS SELECT
stochasticLogisticRegression(0.1, 0.0, 5, 'SGD')(target, x1, x2)
AS state FROM train_data;

Here, we also need to insert data into the train_data table. The number of parameters is not fixed, it depends only on the number of arguments passed into logisticRegressionState. They all must be numeric values. Note that the column with target value (which we would like to learn to predict) is inserted as the first argument.

Predicted labels have to be in [-1, 1].

Predicting

Using saved state we can predict the probability of an object having label 1.

WITH (SELECT state FROM your_model) AS model SELECT
evalMLMethod(model, param1, param2) FROM test_data

The query will return a column of probabilities. Note that first argument of evalMLMethod is an AggregateFunctionState object, next are columns of features.

We can also set a bound of probability, which assigns elements to different labels.

SELECT result < 1.1 AND result > 0.5 FROM
(WITH (SELECT state FROM your_model) AS model SELECT
evalMLMethod(model, param1, param2) AS result FROM test_data)

Then the result will be labels.

test_data is a table like train_data but may not contain target value.

Syntax

stochasticLogisticRegression([learning_rate, l2_regularization_coef, mini_batch_size, method])(target, x1, x2, ...)

Arguments

learning_rate — The coefficient on step length when gradient descent step is performed. A learning rate that is too big may cause infinite weights of the model. Default is 0.00001. Float64
l2_regularization_coef — L2 regularization coefficient which may help to prevent overfitting. Default is 0.1. Float64
mini_batch_size — Sets the number of elements which gradients will be computed and summed to perform one step of gradient descent. Pure stochastic descent uses one element, however having small batches (about 10 elements) makes gradient steps more stable. Default is 15. UInt64
method — Method for updating weights: Adam (by default), SGD, Momentum, Nesterov. Momentum and Nesterov require a little bit more computations and memory, however they happen to be useful in terms of speed of convergence and stability of stochastic gradient methods. String
target — Target binary classification labels. Must be in range [-1, 1]. Float
x1, x2, ... — Feature values (independent variables). All must be numeric. Float

Returned value

Returns the trained logistic regression model weights. Use evalMLMethod for predictions which returns probabilities of object having label 1. Array(Float64)

Examples

Training a model

CREATE TABLE your_model
ENGINE = MergeTree
ORDER BY tuple()
AS SELECT
stochasticLogisticRegressionState(1.0, 1.0, 10, 'SGD')(target, x1, x2)
AS state FROM train_data

Saves trained model state to table

Making predictions

WITH (SELECT state FROM your_model) AS model
SELECT
evalMLMethod(model, x1, x2)
FROM test_data

Returns probability values for test data

Classification with threshold

SELECT result < 1.1 AND result > 0.5
FROM (
WITH (SELECT state FROM your_model) AS model SELECT
evalMLMethod(model, x1, x2) AS result FROM test_data)

Returns binary classification labels using probability threshold

See Also

stochasticLogisticRegression​

stochasticLogisticRegression