Your Browser is Out of Date

Nytro.ai uses technology that works best in other browsers.
For a full experience use one of the browsers below
Demand Forecasting: A Machine Learning-Based Solution to Forecast Demand Signals for Dell Services
Introduction Methodology
Methodology
Demand Forecasting solution components
Impact of insights
Key takeaways Conclusion References

Methodology

Methodology

  • The objective of the Dell Services teams (planning and strategy teams, finance and operation teams, and others) is to efficiently forecast demand signals. These include the following forecasts:

    • Ship
    • Contracts
    • Active Service Units (ASUs)
    • Service Requests (SRs)
    • Dispatches

    The applied data science and engineering team at Dell Technologies created the forecast of different demand singnal by employing ML algorithms that leverage ARIMA time-series demand forecasting model. Auto Regressive Integrated Moving Average (ARIMA) is a class of models that help explain a given time series based on its past values or predicts the future demand on weekly basis, considering all associated attributes such as geographic and product information, entitlements, sales channels, dispatches fields, SRs fields, and more.

    To process huge historical datsets in Hadoop enviornment, Hive and Impala scripts have been leveraged and a timeseries forecasting model is scripted on R and Python. The overall process workflow is automated on SSIS packages (SQL Server Integration Services) and Apache Airflow tool and the result is published into cube for business team review.