liveBook · Manning

Demand prediction

As explained in Math and Architectures of Deep Learning

The task of estimating future product demand from historical and contextual variables.

Demand prediction estimates upcoming product sales from historical and contextual variables. In an e-commerce example, the input vector contains previous-week sales, discount fraction, and weeks until the next holiday; the target is next week's mobile-phone sales. This is framed as a regression problem for inventory management. section 8.5

A PyTorch workflow connects the prediction problem to neural-network training:

  • Represent historical examples as an N × 3 feature matrix and store actual sales in a target vector. section 8.5

How Math and Architectures of Deep Learning covers it

How deep it goes
Connects data representation, normalization, network architecture, loss, and training, while deferring mini-batches to later treatment.
Examples
Uses mobile-phone sales, discounts, holiday timing, a sample dataset, and a PyTorch training loop.
What it assumes
Uses historical feature–target data and introduces the regression setup through a concrete sales-prediction scenario.
What you can do after
Apply demand prediction via PyTorch by preparing data, defining a network, computing loss, and running optimization.
How it connects
Connects regression with neural networks, normalization, mean squared error, stochastic gradient descent, and backpropagation.
Cover of Math and Architectures of Deep Learning
liveBookmanning.com
Read Math and Architectures of Deep Learning on liveBook Get the book at manning.com

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Math and Architectures of Deep Learning. The section links point to the parts of the book it draws on.