5. MINIROCKET

Dempster et al. https://arxiv.org/abs/2012.08791

5.1. Importing Packages

from mlots.models import RidgeClassifierCV
from mlots.transformation import MINIROCKET
from scipy.io import arff
import numpy as np
import warnings
warnings.filterwarnings("ignore")
from sklearn.metrics import accuracy_score

5.2. Loading Data

Here we are loading the SmoothSubspace dataset.

The datasets are in two .arff files with pre-defined train and test splits.

The following code reads the two files stores the X (time-series data) and y (labels), into their specific train and test sets. ***

name = "SmoothSubspace"

dataset = arff.loadarff(f'input/{name}/{name}_TRAIN.arff'.format(name=name))[0]
X_train = np.array(dataset.tolist(), dtype=np.float32)
y_train = X_train[: , -1]
X_train = X_train[:, :-1]

dataset = arff.loadarff(f'input/{name}/{name}_TEST.arff'.format(name=name))[0]
X_test = np.array(dataset.tolist(), dtype=np.float32)
y_test = X_test[: , -1]
X_test = X_test[:, :-1]

#Converting target from bytes to integer
y_train = [int.from_bytes(el, "little") for el in y_train]
y_test = [int.from_bytes(el, "little") for el in y_test]
X_train.shape, X_test.shape

((150, 15), (150, 15))

Set	Sample size	TS length
Train	150	15
Test	150	15

5.3. Transforming Data using MINIROCKET

print("Shape of X_train and X_test before transformation: ",X_train.shape,", ",X_test.shape)

Shape of X_train and X_test before transformation:  (150, 15) ,  (150, 15)

#ts_type denotes if we are using univariate or multivariate version of the algorithm
#we use "univariate" version as the dataset is a univariate time-series
minirocket = MINIROCKET(ts_type="univariate")

minirocket.fit(X_train)
X_train = minirocket.transform(X_train)
X_test = minirocket.transform(X_test)

print("Shape of X_train and X_test after transformation: ",X_train.shape,", ",X_test.shape)

Shape of X_train and X_test after transformation:  (150, 9996) ,  (150, 9996)

5.4. Classification

We can employ RidgeClassifierCV as our linear model for the classification task.

***

model = RidgeClassifierCV(alphas=np.logspace(-3, 3, 10), normalize=True)
model = model.fit(X_train, y_train)

acc = model.score(X_test, y_test)
print(f"Model accuracy: {acc:.2f}%")

Model accuracy: 0.95%