Figure 2
The conceptual workflow shows how raw data is transformed into a graph structure based on relationships between features or nodes. On the left side, a box labeled “Data” contains a simple line plot representing an input signal or time-series data. Below it, an arrow points downward to a small network diagram composed of circular nodes connected by lines, indicating an initial graph representation derived from the data. In the center, a large rounded box illustrates how relationships between nodes are computed. Inside this box, two elements labeled “A subscript i” and “A subscript j” represent two nodes or features. A function is defined below them as “A subscript i, j equals f (A subscript i, A subscript j)”. To the right, two possible outcomes are shown: when “A subscript i, j equals 0”, the nodes “A subscript i” and “A subscript j” are displayed without a connecting line, indicating no edge between them; when “A subscript i, j equals 1”, the nodes are connected by a line, indicating the presence of an edge. On the far right, an arrow points to a more complex network graph with multiple nodes and connections.Construction of time series graph structure. Source(s): Figure created by authors