API Reference - Models - FactorizedPairwiseInteraction
FactorizedPairwiseInteraction is a supervised machine learning model that predicts continuous values (e.g. 1.2, -32, 90, -1.2 and etc. ). It uses iterative calculations to find the best model parameters.
Stored Model Parameters
Contains a matrix.
- ModelParameters[I][J]: Value of matrix at row I and column J. The rows are the features.
Constructors
new()
Create new model object. If any of the arguments are nil, default argument values for that argument will be used.
FactorizedPairwiseInteraction.new(maximumNumberOfIterations: integer, learningRate: number, binaryFunction: string, costFunction: string): ModelObject
Parameters:
-
maximumNumberOfIterations: How many times should the model needed to be trained.
-
learningRate: The speed at which the model learns. Recommended that the value is set between 0 to 1.
-
binaryFunction: The binary function to be used by the model. Available options are:
| Function | Output Range | Skewness | Use Cases |
|---|---|---|---|
| None (Default) | (-∞, ∞) | Symmetric | Rating Prediction |
| Logistic | (0, 1) | Symmetric | Player Choice (A/B), Engagement Prediction, Click-Through Rates |
| HardSigmoid | (0, 1) | Symmetric | Same As Logistic, But Mobile / Real-Time Prediction |
| Probit | (0, 1) | Symmetric | Skill-Based Success, Ability Checks, Normally Distributed Traits |
| BipolarSigmoid | (-1, 1) | Symmetric | Win / Lose, Accept / Reject, Binary Outcomes With Magnitude |
| Tanh | (-1, 1) | Symmetric | Like / Dislike, Positive / Negative Feedback, Preference Modeling |
| SoftSign | (-1, 1) | Symmetric | Gradual Preference Changes, Soft Decisions |
| ArcTangent | (-π/2, π/2) | Symmetric | Academic / Research Alternative To Tanh |
| ComplementaryLogLog | (0, 1) | Right-Skewed | Rare Events Prediction: In-App Purchases, Time-To-Leave Prediction |
| LogLog | (0, 1) | Left-Skewed | Common Events Prediction: Tutorial Completion, Early Wins, First Purchases |
-
costFunction: The function to calculate the cost of each training. Available options are:
-
MeanSquaredError (Default)
-
MeanAbsoluteError
-
BinaryCrossEntropy
-
HingeLoss
-
Returns:
- ModelObject: The generated model object.
Functions
setOptimizer()
Set optimizer for the model by inputting the optimizer object.
FactorizedPairwiseInteraction:setOptimizer(Optimizer: OptimizerObject)
Parameters:
- Optimizer: The optimizer object to be used.
setRegularizer()
Set a regularization for the model by inputting the optimizer object.
FactorizedPairwiseInteraction:setRegularizer(Regularizer: RegularizerObject)
Parameters:
- setRegularizer: The regularizer to be used.
train()
Train the model.
FactorizedPairwiseInteraction:train(featureMatrix: Matrix, labelVector: Matrix): number[]
Parameters:
-
featureMatrix: Matrix containing all data.
-
labelVector: A (n x 1) matrix containing values related to featureMatrix.
Returns:
- costArray: An array containing cost values.
predict()
Predict the value for a given data.
FactorizedPairwiseInteraction:predict(featureMatrix: Matrix): Matrix
Parameters:
- featureMatrix: Matrix containing data.
Returns:
- predictedVector: A vector containing values that are predicted by the model.