Trend analysis is the widespread practice of collecting information and attempting to detect a pattern. It can find the time dependent increase or decrease trends of the variables in your data set and give an alarm according to many parameters you set.
Service can be used to find anomalies in preventive/predictive maintenance and IoT projects and to report these situations in advance.
While scaling the servers in a cloud environment, the usage trend of the servers can be analyzed considering the parameters (disk, network I/O etc.) and adding new servers to the system can be realized automatically.
Our condition-based predictive maintenance product predicts the occurrence of a fault by monitoring the equipment instantaneously and continuously and analyzing it with data of normal operating conditions obtained in the past.
It supports machine learning-based algorithms to predict Remaining Useful Life (RUL).
Depending on data the product is able to work time-series based analysis.
Optimized stock management is essential to keep production running. An optimal holding policy for spare parts helps to optimize the availability of asset generation at a reduced cost
Materials are clustered according to their characteristics such as lead time, unit prices and consumption frequency.
The product is able to find the optimum stock amounts with the quartiles method based on consumption histories.
It also supports traditional heuristics.
Customer segmentation is a marketing method that divides the customers in sub-groups, that share similar characteristics
The product supports clustering method using both state-of-the-art unsupervised learning algorithms (K-Means, DBSCAN, Hierarchical Clustering etc.) and classical methods such as RFM.
Demographic, psychographic, geographical, behavioral clusters are supported
Time Series Analysis (Trend, Seasonality, Irregularity)
Entropy Analysis (Sample & Approximately)
Forecasting with Deep Learning Models (LSTM and GRU)
Forecasting with Statistical Algorithms (ARIMA & SARIMA)
The system can be trained using thousands of product images.
Deep learning models like convolutional neural network (CNN) and its derivatives are used for training.
Image pre-processing tasks are supported if needed.