Forecasting with neural prophet

Time series forecasting powered by neural networks

Published: 2025-12-20 By: Predictiv

Time series forecasting powered by neural networks. This guide covers neuralprophet algorithm overview, training data requirements, seasonality and trend detection, and more.

NeuralProphet algorithm overview

GL Anomaly Detection

The GL Anomaly Detection feature uses a machine learning algorithm to calculate predictions of errors in posted GL transactions. The algorithm uses a machine learning model—a file trained to recognize certain patterns.

Training data requirements

png)

Mobile Stock Take

The Predictiv Stock Take module introduces advanced capabilities for stock take within the Predictiv ERP system, leveraging mobile technology to enhance data collection and streamline stock take processes. Key features include:

  • Mobile Data Collection Integration

  • Verification

  • Advanced Recount and Consolidation Options

  • Workflow-Driven Process

  • Physical Inventory Adjustments

  • Comprehensive Audit Trail

------------------------------------------------------------------------

# AI & Machine Learning

AI Invoice Processing

Supplier invoices are often received in the form of an email containing a PDF attachment. - RPA reads an email account, and extract PDF supplier invoices

  • AI reads the PDF, and extract the invoice details

  • RPA logs in to a Predictiv instance, and captures the invoice details

  • RPA sends an email to the Finance function, alerting the team that there is an invoice which has been captured, and should be validated and complete

Sales Forecasts

Predictiv supports a 'pluggable' sales forecasting capability, whereby future demand can be estimated using historical data.

Seasonality and trend detection

Sales forecasting uses a decomposable time series model with the components, trend, seasonality, auto-regression, special events, future regressors and lagged regressors. ## GL Anomaly Detection

The GL Anomaly Detection feature uses a machine learning algorithm to calculate predictions of errors in posted GL transactions.

Forecast accuracy metrics

  • RPA reads an email account, and extract PDF supplier invoices

  • AI reads the PDF, and extract the invoice details

  • RPA logs in to a Predictiv instance, and captures the invoice details

  • RPA sends an email to the Finance function, alerting the team that there is an invoice which has been captured, and should be validated and complete

Sales Forecasts

Predictiv supports a 'pluggable' sales forecasting capability, whereby future demand can be estimated using historical data. Sales forecasting uses a decomposable time series model with the components, trend, seasonality, auto-regression, special events, future regressors and lagged regressors. Forecasts can be run on a daily schedule, with a configurable forecast window.

Integration with demand planning

**

------------------------------------------------------------------------

# Predictiv Mobile

The Predictiv Mobile app represents a major addition to the ways in which users can engage with Predictiv. png)

Mobile Stock Take

The Predictiv Stock Take module introduces advanced capabilities for stock take within the Predictiv ERP system, leveraging mobile technology to enhance data collection and streamline stock take processes. Key features include:

  • Mobile Data Collection Integration

  • Verification

  • Advanced Recount and Consolidation Options

  • Workflow-Driven Process

  • Physical Inventory Adjustments

  • Comprehensive Audit Trail

------------------------------------------------------------------------

# AI & Machine Learning

AI Invoice Processing

Supplier invoices are often received in the form of an email containing a PDF attachment. - RPA reads an email account, and extract PDF supplier invoices

  • AI reads the PDF, and extract the invoice details

  • RPA logs in to a Predictiv instance, and captures the invoice details

  • RPA sends an email to the Finance function, alerting the team that there is an invoice which has been captured, and should be validated and complete

Sales Forecasts

Predictiv supports a 'pluggable' sales forecasting capability, whereby future demand can be estimated using historical data. Sales forecasting uses a decomposable time series model with the components, trend, seasonality, auto-regression, special events, future regressors and lagged regressors.

Getting Started

To implement neuralprophet sales forecasting in your Predictiv environment:

1. Assess your current state - Review existing processes and identify improvement opportunities

2. Configure the module - Work with your implementation team to set up the required components

3. Train your team - Ensure users understand the new capabilities and workflows

4. Monitor and optimize - Track key metrics and continuously improve

Related Resources

For more information on related topics, explore our other guides in this collection.

Need Help?

Our team of experts is available to help you get the most out of Predictiv. Contact us to discuss your specific requirements and how we can help you achieve your goals.

Need Help Implementing This?

Our team can help you put these best practices into action.