PHP Machine Learning Rain Forecaster is a simple machine learning experiment in predicting rain based on a few forecast indicators.

PHP Machine Learning Rain Forecaster is a simple machine learning experiment in predicting rain based on a few forecast indicators.: forecasted "HighTemp", "LowTemp", "AverageTemp", "DewPoint", "DewPointHigh", "DewPointLow", "DewPointAverage", and forecasted "SeaLevelPressure".

It's just an experiment in Machine Learning. DewPoint seems critical and a great indicator.

I am using the excellent PHP machine learning library for PHP, Rubix ML: https://rubixml.com/

Just when you thought it was only OK to do machine learning in Python or Java :-)

Samples in weather.csv were built from https://www.wunderground.com/history for Portland, Oregon, 1st of every month for 2020

REQUIREMENTS

PHP ^7.4

Composer

INSTALLATION

1. Clone the repo
2. Run "composer install"

TESTING / USAGE: machine_learning_rain_prediction.php command line script

Test:

See the predictions from the weather.csv samples and the unknown samples in machine_learning_rain_prediction.php. From the command line, run:

php machine_learning_rain_prediction.php

Suggestions for your learning:

1. Play with the data in weather.csv to create your own samples. Head to https://www.wunderground.com/history and build a sample set for your city.
2. Build an unknown.csv and load that instead of the hard coded unknown samples in machine_learning_rain_prediction.php.

Have fun!