Case study
Predictive Analytics
Cost Forecasts
The Challenge
Construction projects regularly exceed calculated budgets.
By leveraging historical data it is possible to radically improve the accuracy of price forecasts.
The Solution
COMBINING MACHINE LEARNING WITH HUMAN EXPERTISE
1. Input parameter, such as project size and product type are analyzed and matched with a predefined ruleset.
2. A classification model trained 3.5 on historical data identifies necessary cost items.
3. A bayesian algorithm learns the relation between parameter and cost items while accurately modeling uncertainty
Benefits
- Better decision making based on improved forecast accuracy
- Transparent forecasts by modeling cost uncertainties
- Automated calculations speed up the proposal process
- Continuous improvements by using online learning and human feedback
Further Use Cases
- Real Estate: Calculating projected home values
- Healthcare: Predicting the effectiveness of a treatment
- Advertising: Finding new market segments or forecast the success of activities
- Customer Service: Predicting your demand and staff need
... and many more
Other Case Studies
Contact us