Data Science
Table of Contents:
AnomalyDetection
A data science component to apply different anomaly detection models. The component only needs a data source to apply the model on.
The anomaly detection models only return a list of anomalies. ChartData2D component is needed to highlight the anomalies on a chart
Properties
None
Events
None
Methods
- CleanData(anomaly,xList,yList)
- Given a single anomaly: [(anomaly index, anomaly value)]
- Iterate over the xList and delete value at anomaly index
- Iterate over the yList and delete the value at anomaly index with the same value as anomaly value
- combine the xList and yList after modification in a list of x and y pairs
We assume x and y lists are the same size and are ordered.
- DetectAnomalies(dataList,threshold)
- Calculates the mean and standard deviation of the data, and then checks each data point’s Z-score against the threshold. If a data point’s Z-score is greater than the threshold, the data point is labeled as anomaly.
- DetectAnomaliesInChartData(chartData,threshold)
- Detects anomalies in the given chart data object by comparing Y values to the threshold based on their standard deviation to the mean.
Regression
A data science component to apply different regression models. The component only requires a data source to apply the model on.
The component is only responsible for the statistical calculations and provides the following properties for line of best fit: “slope”, “Yintercept”, “correlation coefficient”, and “predictions”
To draw the line of best fit use the drawing block in ChartData2D component
Properties
None
Events
None
Methods
- CalculateLineOfBestFitValue(xList,yList,value)
- Returns one of the Line of Best Fit values. A value could be “slope”, “Yintercept”, “correlation coefficient”, “predictions” or a dictionary when AllValues is specified.