Which solution will meet these requirements?

By:
On:
In: MLS-C01
Tagged:
With: 0 Comments
A retail company wants to use Amazon Forecast to predict daily stock levels of inventory. The cost of running out of items in stock is much higher for the company than the cost of having excess inventory. The company has millions of data samples for multiple years for thousands of items. The company’s purchasing department needs to predict demand for 30-day cycles for each item to ensure that restocking occurs.A machine learning (ML) specialist wants to use item-related features such as “category,” “brand,” and “safety stock count.” The ML specialist also wants to use a binary time series feature that has “promotion applied?” as its name. Future promotion information is available only for the next 5 days.The ML specialist must choose an algorithm and an evaluation metric for a solution to produce prediction results that will maximize company profit.

Which solution will meet these requirements?

Train a model by using the Autoregressive Integrated Moving Average (ARIMA) algorithm. Evaluate the model by using the Weighted Quantile Loss (wQL) metric at 0.75 (P75).

Train a model by using the Autoregressive Integrated Moving Average (ARIMA) algorithm. Evaluate the model by using the Weighted Absolute Percentage Error (WAPE) metric.

Train a model by using the Convolutional Neural Network – Quantile Regression (CNN-QR) algorithm. Evaluate the model by using the Weighted Quantile Loss (wQL) metric at 0.75 (P75).

Train a model by using the Convolutional Neural Network – Quantile Regression (CNN-QR) algorithm. Evaluate the model by using the Weighted Absolute Percentage Error (WAPE) metric.

Explanations:

The ARIMA algorithm is not suitable for incorporating item-related features and time series binary features like “promotion applied?” which are necessary for this problem. Additionally, ARIMA does not handle the 30-day prediction cycle or the complex dependencies between features well.

Similar to Option A, ARIMA is not appropriate for this use case due to its limitations with complex features like binary promotion data and item-related features. WAPE is also not the best metric for this scenario, as it doesn’t consider quantile losses, which are important in predicting stock levels.

The CNN-QR algorithm is a good choice because it is capable of handling multiple input features, including item-related and binary time series features. The Weighted Quantile Loss (wQL) metric at 0.75 is appropriate for this problem as it focuses on minimizing the error for higher demand forecasts, which is crucial when managing stock levels to avoid running out of inventory.

While the CNN-QR algorithm is a good choice, WAPE is not the ideal metric for this situation. The problem’s focus on minimizing stockouts and maximizing profit requires a more nuanced metric like wQL, which is better suited for forecasting demand at specific quantiles. WAPE does not provide the necessary focus on higher-demand scenarios.

Leave a Reply

Your email address will not be published. Required fields are marked *

5 × one =