The world of horse racing is dynamic and ever-changing, with countless factors influencing the outcome of a race. In this competition, we delve deep into the nuances of horse speeds, presenting a challenge that requires participants to predict the speed of a horse during a race.

The provided dataset encapsulates over 100,000 races, each with its unique set of conditions. The primary variables include the date of the race, the time the race took place, and the weather condition during the race. The challenge is to use this data to predict the speed of a horse under different conditions.


Dataset
The dataset is divided into training and testing sets. The training set consists of 100,000 rows, each row representing a unique race. The columns in the dataset are:

Date: The date the race took place.

Race_Time: The time of the day the race occurred.

Weather: The weather condition during the race. It has categories such as 'Sunny', 'Cloudy', 'Rainy', etc.

Horse_Speed: The speed of the horse during the race. This is the target variable for the training set.
The testing set follows a similar structure but lacks the Horse_Speed column. 


Participants are expected to predict this missing column using the given conditions.

Evaluation


The Mean Absolute Error (MAE) is used to evaluate submissions in this competition. It measures the average magnitude of errors between predicted and actual horse speeds. It's calculated as:

MAE = (1/n) * SUM(|y_i - yhat_i|)

Where:

n is the number of observations.
y_i is the actual horse speed for the ith race.
yhat_i is the predicted horse speed for the ith race.
Aim to minimize the MAE for a higher leaderboard position. The closer the MAE is to zero, the better the predictions





​

Challenge
The intricate patterns and relationships between the variables in this dataset make it an elite-level challenge. While the variables might seem simple at first glance, their combined interactions create a complex landscape for prediction. This competition is designed to push the boundaries of traditional machine learning and encourage innovative approaches.

Are you ready to take on the challenge and gallop your way to victory?