HPA has developed numerous predictive analysis projects using the leading statistical and machine learning techniques and models, including:
- Linear and logistic regression
- Decision trees
- Random ForestSVM (Support Vector Machine)
- Naïve Bayes
- KNN (k-Nearest Neighbors)
- Neural Network (NN)
- XGBoost
Here are some applications developed:
#ENERGY
- Forecasting of electricity production and consumption
- Trading on electricity markets (CCC auctions)
- Reconstruction of daily and multi-day consumption curves
- Consumption plausibility (water, gas, electricity)
- Predictive analysis for the remaining battery life of smart-meters
#MANUFACTURING
- Predictive maintenance and fault detection within the production line
#ENVIRONMENT AND #WASTEMANAGEMENT
- Prediction of waste containers filling
#SECURITY
- Automatic classification of environmental sensors alerts
#CRM
- Churn prediction and next best action
- Ticket classification and first response
