Skip to content

Linear Regression SGD Model

A linear regression model assumes that there is a linear relationship between the input variables X and the target/output value y. In practice this implies that there exists some linear combination of the input variable X which allows us to infer/predict the value y.

This relationship can be described by the formula:

\[y_i= \sum_{n=1}^{N} \theta_{n} X_{i,n}\]

Where:

Parameter Description
X Are the N dimensional features which in linear combination represent y
y Is the variable being predicted
\(\theta\) Are the weighted coefficients used to calculate the best linear combination of X to describe y

The goal of a linear regresion model is to calculate the theta coefficients which can then be used in combination with input data to make predictions based on this linear relationship.

.ml.online.sgd.linearRegression.fit

Fit a linear regression model using stochastic gradient descent

.ml.online.sgd.linearRegression.fit[X;y;trend;paramDict]

Parameters:

name type description
X any Input/training data of N dimensions.
y any Output/target regression data.
trend boolean Is a trend to be accounted for.
paramDict dictionary Any modifications to be applied during the fitting process of SGD (See here for more details).

Returns:

type description
dictionary All information collected during the fitting of a model, along with prediction and update functionality. updateSecure has also been included to allow new data to be used to update the model where additional checks are applied to the data to ensure that it is in the correct format to ensure no 'model pollution' occurs.

The information collected during the fitting of the model are contained within the modelInfo key and includes:

name description
theta The weights calculated during the process.
iter The number of iterations applied during the process.
diff The difference between the final theta values and the preceding values.
trend Whether or not a trend value was fitted during the process.
paramDict The parameter dictionary used during the process.
inputType The data type of each column of the input data.

Prediction functionality is contained within the predict key. The function takes the following inputs:

  • X is the input/training data of N dimensions

and returns the predicted values

The model contains two types of update functions

  • update, where models are updated assuming that the data given is suitable
  • updateSecure, where additional checks are applied to the data to ensure that it is in the correct format to ensure no 'model pollution' occurs.

Both functions take the following inputs:

  • X is the input/training data of N dimensions
  • y is the output/target regression data

returns a dictionary containing all information collected during the updating of a model, along with a prediction and update function.

If updateSecure is used, an error will be returned if appropriate data is not used. See here for more information.

During the update phase, the same model parameters are used that were applied during the fitting process, except the maximum iteration is set to 1.

Examples:

Example 1: Fit, predict and update a model

// Create data with strong correlation but also some noise
q)X:8*100?1f
q)y:4+3*X+100?1f

// Fit a linear regression SGD
q)show regMdl:.ml.online.sgd.linearRegression.fit[X;y;1b;`maxIter`alpha!(1000;0.01)]
modelInfo   | `theta`iter`diff`trend`paramDict`inputType!(5.648133 2.934584;1000;..
predict     | {[config;X]
  config:config`modelInfo;
  theta:config`theta;
  tre..
update      | {[config;X;y]
  modelInfo:config`modelInfo;
  theta:modelInfo`thet..
updateSecure| {[config;secure;X;y]
  modelInfo:config`modelInfo;
  theta:mode..

// Information generated during the fitting of the model
q)regMdl.modelInfo
theta    | 5.648133 2.934584
iter     | 1000
diff     | 0.01827465 0.01503591
trend    | 1b
paramDict| `alpha`maxIter`gTol`theta`k`seed`batchType`gra...
inputTyp | -9h

// Predict on new data
q)Xnew:8*10?1f
q)regMdl.predict[Xnew]
12.26865 19.97765 24.66665 18.05956 13.65026...

// Update the model using new data points
q)Xupd:8*5?1f
q)yUpd:4+3*Xupd+5?1f
q)show regUpd:regMdl.update[Xupd;yUpd]
modelInfo   | `theta`iter`diff`trend`paramDict!(5.627178 2.926986;1;0.02095537 0..
predict     | {[config;X]
  config:config`modelInfo;
  theta:config`theta;
  tre..
update      | {[config;X;y]
  modelInfo:config`modelInfo;
  theta:modelInfo`thet..
updateSecure| {[config;secure;X;y]
  modelInfo:config`modelInfo;
  theta:mode..

// Information generated during the updating of the model
q)regUpd.modelInfo
theta    | 5.627178 2.926986
iter     | 1
diff     | 0.02095537 0.007598042
trend    | 1b
paramDict| `alpha`maxIter`gTol`theta`k`seed`batchType`gradArgs`penalty`lambda.
inputTyp | -9h

Configurable parameters

In the above function, the following are the optional configurable entries for paramDict:

name type description default
alpha float Learning rate applied. 0.01
maxIter integer Max possible number of iterations before the run is terminated, this does not guarantee convergence. 100
gTol float If the difference in gradient falls below this value the run is terminated. 1e-5
theta float Initial starting weights. 0
k integer Number of batches used or random points chosen each iteration. *n
seed integer Random seed. random
batchType symbol Batch type - `single`shuffle`shuffleRep`nonShuffle`noBatch. shuffle
penalty symbol Penalty/regularization term - `l1`l2`elasticNet. l2
lambda float Penalty term coefficient. 0.001
l1Ratio float Elastic net mixing parameter - only used if penalty type is ElasticNet is applied. 0.5
decay float Decay coefficient. 0
p float Momentum coefficient. 0
verbose boolean If information about the fitting process is to be printed after every epoch. 0b
accumulation boolean If the theta value after each epoch is returned as the output 0b
thresholdFunc list Threshold function and value to apply when using updateSecure ()

In the above table *n is the length of the dataset.

A number of batchTypes can be applied when fitting a model using SGD, the supported types and an explanation of their use of the k parameter are explained below:

options:

name description
noBatch No batching occurs and all data points are used (regular gradient descent)
nonShuffle Data split into k batches with no shuffling applied.
shuffle Data shuffled into k batches. Each data point appears once.
shuffleRep Data shuffled into k batches. Data points can appear more than once and not all data points may be used.
single k random points are chosen each iteration.