Predictive Model

The Predictive Model component is a tool within the Nappai dashboard that allows you to get predictions and recommendations from your data. It acts as a bridge to the AI engine, analyzing information to return results like predicted values, confidence scores, or tailored suggestions. You can use this component to automate decisions based on data insights without needing to write code or manage complex configurations.

How it Works

Think of this component as a dedicated assistant for making predictions. Inside Nappai, the heavy lifting of calculations and AI logic is handled by a powerful background engine. This component is your interface to that engine.

When you place this component in your workflow, you are essentially telling Nappai to analyze the data you provide. The component takes your inputs, runs them through the prediction algorithms, and outputs the results. These results might include:

• Predicted Values: Numerical forecasts based on trends.
• Confidence Scores: Indicators of how certain the prediction is.
• Recommendations: Lists of suggested actions or items.

The component ensures that the data is processed efficiently and securely. It uses standard helpers to manage the information, ensuring that your workflow remains stable and easy to maintain.

Inputs

Input Fields

This component uses input fields that are shared with the underlying prediction engine. The specific fields you see will depend on how the prediction model is configured in your system.

• Dynamic Configuration: The available inputs may vary based on the type of prediction you are running.
• Data Requirements: Ensure you provide the necessary data fields required by the model to generate accurate predictions.

Outputs

This component produces results that you can use in subsequent steps of your workflow. The output typically includes structured data such as prediction values, scores, and recommendations.

Output Data Example (JSON)json

{ “predictions”: [105.5, 110.2, 98.7], “confidence_scores”: [0.92, 0.88, 0.95], “recommendations”: [“Action A”, “Action B”], “metadata”: { “model_type”: “PredictiveModel”, “status”: “success” } } You can map these outputs to other components to trigger actions, update records, or display insights in your dashboard.

Connectivity

This component is designed to fit seamlessly into automation workflows. It typically connects to nodes that consume prediction results, such as:

• Action Nodes: To trigger tasks based on predicted values.
• Reporting Dashboards: To visualize predictions and confidence scores.
• Decision Branches: To route workflows based on recommendations or thresholds.

Connecting this component to downstream nodes allows you to automate responses based on the insights generated by the model.

Usage Example

Scenario: Automating inventory replenishment based on sales predictions.

1. Add the Predictive Model component to your workflow.
2. Connect your sales data to the component’s input fields.
3. The component processes the data and outputs predicted sales numbers along with confidence scores.
4. Map the predicted sales numbers to a trigger that automatically creates purchase orders.
5. Use the confidence scores to decide whether to send a high-priority alert to the management team.

Tips and Best Practices

• Data Quality: Ensure your input data is clean and accurate to get the best predictions.
• Workflow Design: Use this component when you need AI-driven forecasts or recommendations.
• Stability: The component is designed for reuse and stability; you can rely on it to provide consistent outputs as long as inputs are valid.
• Non-Technical Focus: You do not need to understand the internal algorithms. Simply provide the data and use the results.

Security Considerations

• Data Privacy: Ensure that the data you feed into the Predictive Model component complies with your organization’s data privacy policies.
• Secure Processing: The component processes data securely within the Nappai framework, adhering to system standards for data handling.