Machine Learning

Please enter TSMLTC (Time Series Machine Learning Test Center) here.

Provided are services for creating and maintaining arbitrary time series for testing purposes. For each series you can calculate a regression along with its quadratic deviation from the series as well as a forecast for the next point in time.

Moreover, you can configure and execute training of a neural network. Learning rate, training iterations and size of training data determine training results. Last but not least, quadratic deviation from the series and forecast of the next point in time are generated for performance comparison.

Neural networks are complex data structures designed to learn from supervision. Training data is structured in such a way that each stimulus or input comes with a target or output to be learned by supervision. Once neural networks are trained they infer targets from given inputs. In context of time series past targets are learned from historical data for predicting future output.