☰
The hardest data science tournament on the planet.
2,200 models are building the world's open hedge fund by predicting the stock market.
Start with hedge fund quality data.
It is
clean and regularized, designed to be usable right away.
Obfuscated, so it can be given out for free.
Download Dataset
| id | era | feature1 | . . . | feature310 | target |
|---|---|---|---|---|---|
| n2b2e3dd163cb422 | era1 | 0.75 | . . . | 0.00 | 0.25 |
| n177021a571c94c8 | era1 | 1.00 | . . . | 0.25 | 0.75 |
| n7830fa4c0cd8466 | era1 | 0.25 | . . . | 1.00 | 0.00 |
| nc584a184cee941b | era1 | 0.25 | . . . | 0.00 | 1.00 |
| nc5ab8667901946a | era1 | 0.75 | . . . | 0.25 | 0.25 |
| n84e624e4714a7ca | era1 | 0.00 | . . . | 0.75 | 1.00 |
Apply machine learning to predict the stock market.
Build a model
using the example Python and R scripts. Everything you need to get started in one package.
Clone Example Scripts
#!/usr/bin/env python
""" Example classifier on Numerai data using a xgboost regression. """
import pandas as pd
from xgboost import XGBRegressor
# training data contains features and targets
training_data = pd.read_csv("numerai_training_data.csv").set_index("id")
# tournament data contains features only
tournament_data = pd.read_csv("numerai_tournament_data.csv").set_index("id")
feature_names = [f for f in training_data.columns if "feature" in f]
# train a model to make predictions on tournament data
model = XGBRegressor(max_depth=5, learning_rate=0.01, \
n_estimators=2000, colsample_bytree=0.1)
model.fit(training_data[feature_names], training_data["target"])
# submit predictions to numer.ai
predictions = model.predict(tournament_data[feature_names])
predictions.to_csv("predictions.csv")| id | prediction |
|---|---|
| n60dffdaceb7e467 | 0.25 |
| nadaeef0214b84a8 | 1.00 |
| nb13883520a4344f | 0.25 |
| n423766c5a4fa42a | 0.75 |
| n252b14301e46a31 | 0.25 |
| n75a5baf93a624cc | 0.00 |
| n2ff91086716e413 | 1.00 |
