Gbm

Here, we present the model's performance in predicting the bitcoin price along with their hyper-parameters. For more details on the model, see here.

Prediction

For scoring, Root Mean Squared Error was used.

Mean across all targets: 1,455.17

D+7	D+14	D+21	D+28
1,487.02	1,383.88	540.04	2,409.72

Mean across all targets: 3,309.61

D+7	D+14	D+21	D+28
3,282.33	3,718.84	4,358.90	1,878.39

Below are choices made for the model setup.

(Determined by Forward Selection and heuristically based on the model's assumptions)

param	description	value
include_indices	Whether to include market indices.	False
include_commodities	Whether to include commodities.	True
include_currencies	Whether to include currencies.	False

(Determined by Grid Search)

param	description	value
ohlc_ratio	Whether to add OHLC ratios of bitcoin price to the data.	False
lags	List of lags to use. None, if no lags are used.	[7, 14]

(Determined by heuristically, considering the model's assumptions)

param	description	value
log	Whether to take the log of the data.	False
diff	Whether to difference the data.	False
normalize	Whether to normalize the data.	False

(Determined by Grid Search)

param	description	value
max_depth	The maximum depth of the tree. A negative value means no limit.	10
min_data_in_leaf	The minimum number of data in a leaf. A smaller number may reduce overfitting.	3
n_estimators	The number of trees in the forest.	20
colsample_bytree	The fraction of features to use for each tree.	0.7