Understanding Models Learn about the different types of models available

BorBann uses machine learning models to analyze property data and make predictions. These models are trained on historical property data and can be used to predict property prices, identify trends, and provide insights.

Types of Models

Regression Models Standard ML Model v2.4

Used for predicting continuous values like property prices. These models analyze various features to estimate a property's value.

Neural Networks Enhanced Neural Network v1.8

Deep learning models that can capture complex patterns in property data. Ideal for analyzing multiple factors simultaneously.

Geospatial Models Geospatial Regression v3.1

Specialized models that incorporate location data and spatial relationships between properties.

Time Series Models Time Series Forecast v2.0

Models designed to analyze property price trends over time and make future predictions.

System vs. Custom Models

System Models are pre-trained models provided by BorBann. They are regularly updated and maintained by our team.

Custom Models are models that you train using your own data pipelines. These can be tailored to your specific needs.

Using Models How to select and use models for property analysis

Selecting a Model

You can select different models when using the Maps or Price Prediction features. Look for the model selector dropdown in the interface.

Step-by-Step Guide

1
Navigate to the Maps or Price Prediction page
2
Look for the model selector dropdown in the top navigation bar
3
Click on the dropdown to see available models
4
Select the model that best suits your analysis needs
5
The page will update to use the selected model for analysis

Understanding Model Results

Different models may produce slightly different results. Here's how to interpret them:

Price Predictions

Models provide a predicted price along with a confidence level. The higher the confidence, the more reliable the prediction.

Feature Importance

Models show which features (location, size, etc.) have the most impact on the property price.

Price Range

Models provide a range of possible prices based on the confidence level.

Environmental Impact

Models analyze how environmental factors affect property values in the area.

Training Custom Models Learn how to create and train your own models

You can train custom models using your own data pipelines. This allows you to create models tailored to your specific needs.

Step-by-Step Guide

1

Navigate to the Models page
Go to the Models section from the sidebar navigation
2

Click "Train New Model"
This will take you to the model training interface
3

Configure your model
Enter a name, description, and select the model type
4

Select a data pipeline
Choose which data pipeline to use for training the model
5

Start training
Click the "Start Training" button to begin the training process
6

Monitor training progress
The system will show you the training progress and notify you when it's complete

Important Notes

Training a model requires a data pipeline with sufficient data
The training process may take several minutes depending on the data size
You can cancel training at any time if needed
Models with more data generally produce more accurate results

Quick Links Related Guides

Price Prediction

Learn how to use models for property price prediction

Data Pipelines

Set up data pipelines for model training

Maps & Geospatial

Use models with the interactive map

Need Help?

Can't find what you're looking for? Our support team is here to help.