Intelligent solutions for samrt cities problems.

MCS Data Labs is a German SME specialized in the representation structured and unstructured Big Data. MCS Data Labs support using new analytics techniques to integrate and better visualize the data flowing from the social media and the rescue teams members that are on the emergency scene.

This project is regarding a binary classification using both classic and ensemble machine learning algorithms to detect the fake task.

Sensors are randomly distributed in a environment. They transmit the temperature, humidity and pressure data to the only controller at regular intervals. The controller processes the received data and gives a predicted value to reflect the parameters of the current environment.

Theta is the estimated value of environment parameters (temperature, humidity or pressure).

Network simulation for both normal and attacker behaviour using contiki-ng simulator. Using multi-classification machine learning techniques to distinguis among network behaviour. Naive Bayes classifier has performed the best between random forest and adaboost.

Anomaly detection using a threshold by which we can classify data with certain interval as an anomaly. Idea: we combine the two data frames predicted to be anomaly data construct a data frame of anomalies.
Algorithms used : K-Means and the One-Class Support Vector.

Generating tasks and user movement events using stochastic algorithm.