Why time series has unique challenges in ML Python - Challenge Your Understanding

Challenge - 5 Problems

🧠 Conceptual

intermediate

Why is stationarity important in time series analysis?

In time series data, why do many models require the data to be stationary?

A Because stationary data contains no noise, so models perform better.

B Because stationary data has constant mean and variance over time, making patterns easier to model.

C Because stationary data always has increasing trends that models can learn.

D Because stationary data has random values that change unpredictably.

💡 Hint

Think about what it means for data to have consistent behavior over time.

Stationarity means the data's statistical properties like mean and variance do not change over time. This helps models learn stable patterns and make reliable predictions.

❓ Predict Output

intermediate

Output of lag feature creation in time series

What is the output of the following Python code that creates a lag feature?

ML Python

import pandas as pd

data = pd.Series([10, 20, 30, 40, 50])
lag_1 = data.shift(1)
print(lag_1.tolist())

A [0, 10, 20, 30, 40]

B [10, 20, 30, 40, 50]

C [20.0, 30.0, 40.0, 50.0, None]

D [nan, 10.0, 20.0, 30.0, 40.0]

💡 Hint

Recall that shift(1) moves data down by one position, introducing a missing value at the start.

The shift(1) function moves all values down by one, so the first value becomes nan and the rest are previous values.

❓ Model Choice

advanced

Best model choice for seasonal time series data

You have monthly sales data with clear yearly seasonality. Which model is best suited to capture this pattern?

A ARIMA model with seasonal components (SARIMA)

B Linear Regression without any seasonal terms

C K-Means clustering

D Decision Tree classifier

💡 Hint

Consider models designed to handle seasonality in time series.

SARIMA extends ARIMA by including seasonal terms, making it suitable for data with repeating seasonal patterns.

❓ Metrics

advanced

Choosing the right error metric for time series forecasting

Which error metric is most appropriate when you want to measure the average percentage error in time series forecasts?

A Mean Absolute Error (MAE)

B Root Mean Squared Error (RMSE)

C Mean Absolute Percentage Error (MAPE)

D Accuracy Score

💡 Hint

Think about which metric expresses error as a percentage.

MAPE expresses forecast errors as percentages, making it easy to understand relative error size.

🔧 Debug

expert

Identifying the cause of poor time series model performance

A time series model trained on daily temperature data shows poor predictions on recent data. Which issue is most likely causing this?

A The model was trained on stationary data but recent data has a sudden trend change (concept drift).

B The model used too many lag features causing overfitting.

C The model was trained with too few epochs.

D The model used a wrong activation function.

💡 Hint

Consider what happens when the data pattern changes after training.

Concept drift means the data distribution changes over time, so a model trained on old data may not predict well on new data.