Google Professional Machine Learning Engineer exam dumps

Google Professional Machine Learning Engineer practice question 15 of 522

Professional Machine Learning Engineer. Professional level, Google Cloud. Free question with the correct answer and a full explanation.

Google Professional Machine Learning Engineer Question 15

Single answerGoogle Cloud Platform

You are working on a project to predict product demand for an e-commerce platform. The dataset includes daily sales data for several products over the past 5 years, and you need to build a machine learning model to forecast future demand. After exploratory data analysis, you notice strong seasonal trends and periodic spikes in sales. Which of the following modeling approaches is most appropriate for this task?

  1. A

    Train a boosted tree regression model with the raw sales data as input.

  2. B

    Apply an autoencoder to reduce the dimensionality of the dataset and then use the compressed features for prediction.

  3. C

    Use a time-series model such as ARIMA or a recurrent neural network (RNN) for forecasting.

  4. D

    Use a matrix factorization approach to identify latent features in the sales data and predict demand.

Show answer and explanation

Correct answer: C

Explanation

Time-series forecasting requires models that can effectively capture temporal dependencies, trends, and seasonality in data. ARIMA and RNNs are specifically designed for such tasks, making them the most appropriate choice for forecasting product demand in this scenario. Other approaches, like boosted trees, autoencoders, and matrix factorization, lack the necessary mechanisms to handle the sequential nature and seasonal patterns inherent in the dataset.

  • A. Incorrect.

    Boosted trees are versatile and powerful for regression tasks, but they are not specifically designed to handle sequential dependencies or seasonal patterns in time-series data, which are crucial for accurate forecasting in this scenario.

  • B. Incorrect.

    Autoencoders are primarily used for dimensionality reduction or anomaly detection, but they are not inherently suitable for time-series forecasting tasks as they do not capture temporal dependencies in the data.

  • C. Correct.

    Time-series models like ARIMA or RNNs are explicitly designed to handle sequential data and can capture trends, seasonality, and temporal dependencies, making them the most appropriate choice for forecasting product demand in this scenario.

  • D. Incorrect.

    Matrix factorization is typically used for recommendation systems or collaborative filtering tasks. While it can uncover latent features, it is not well-suited for modeling temporal dependencies in time-series data.

Timed practice exam

Take a Google Professional Machine Learning Engineer practice test under exam conditions

60 questions in 120 minutes, drawn from this bank, with a score report and a per-question review when you finish.

Start timed exam