Neural Network Analysis of Hammer Throw Distance using Scikit-Learn

Jorgen Bergstrom included in Hammer Throw

2024-05-28 250 words 2 minutes

Contents

Summary

The MLPRegressor in scikit-learn is a powerful tool for performing regression tasks using a multi-layer perceptron (MLP), which is a type of artificial neural network. It is a supervised learning algorithm that learns a function that maps input data to continuous output values. It can model complex relationships between features and the target variable.

Code Location

You can find the code in this example here.

import matplotlib.pyplot as plt
import numpy as np
from sklearn.model_selection import train_test_split
from sklearn.neural_network import MLPRegressor
from sklearn.preprocessing import StandardScaler  
from sklearn.metrics import mean_squared_error, mean_absolute_error 

X = np.genfromtxt('data_X.csv', delimiter=',')
Y = np.genfromtxt('data_Y.csv', delimiter=',')

# Split the data into training and testing sets. test_size=0.3 indicates that 30%
# of the data will be used for testing, and 70% for training.
X_train, X_test, y_train, y_test = train_test_split(X, Y, test_size=0.3)
# Transform the input X data by subtracting the mean and dividing with the standard dev.

scaler = StandardScaler()
X_train = scaler.fit_transform(X_train)
X_test = scaler.transform(X_test)

# MLP = multi layer perceptron
# This sets up the whole Neural Network model
mlp = MLPRegressor(hidden_layer_sizes=(50, 25, 12, 6, 3), activation='relu', solver='adam', max_iter=10000, verbose=False)
mlp.fit(X_train, y_train)
predictions = mlp.predict(X_test)

print(f"Mean absolute error={mean_absolute_error(y_test, predictions):.3f}")

plt.scatter(X_test[:,1], y_test, c='r')
plt.scatter(X_test[:,1], predictions, c='b')
plt.xlabel('Velocity (m/s)')
plt.ylabel('Distance (m)')

The output from running this code is: Mean absolute error=0.727

This example shows how simple it can be to set up and run a Neural Network (NN) model. The results show that the model predictions are decent, but not as good as a 4th order polynomial regression model.