ft_linear_regression

It is the simple iOS application for training linear regression with gradient descent optimization algorithm. You can change parameters to see how do they affect learning.

Tested with:

• macOS Catalina - 10.15.5 Beta
• Xcode - 11.4.1
• iPhone 8
• iOS - 13.3.1

Get started:

git clone https://github.com/Gleonett/ft_linear_regression.git
cd ft_linear_regression
git submodule update --init --recursive

1. Open ft_linear_regression/ft_linear_regression.xcodeproj
2. Drag the Charts/Charts.xcodeproj to the project
3. Go to your target's settings, hit the "+" under the "Embedded Binaries" section, and select the Charts.framework
4. Build and run

Explanation

Main view

Main view contains visualization for model, dataset and buttons:

• Parameters - Link to Parameters view
• Reset - Reset already trained linear regression model parameters
• Train - Train model
• Predict - Link to Prediction view

Parameters view

Learning rate is a tuning parameter in an optimization algorithm that determines the step size at each iteration while moving toward a minimum of a loss function

learningRate = initinalLearningRate / (decay * currentEpoch + 1)

Model will train either until the end of epochs or until:

abs(InterceptN - InterceptN-1) < AccuracyThreshold
and
abs(BiasN - BiasN-1) < AccuracyThreshold

Prediction view

regressor is the value we want to predict the dependent