Identifying the right algorithm for a specific problem has been a daunting task lately. This is because different problems present different levels of complexity. It is therefore important to identify the complexity of the problem and to divide the complexity as stated by Herbert Simon in his famous article [1]. Simplification of the process of deciphering the complexity is extremely important. More recently, much of this has been carried out using graphs. This paper identifies a new improvement in algorithmic development using a flow diagram to identify state sequences.

The paper identifies a flow on the set of states that can be recognized by a sensor. For example, it identifies a particular state number when an individual is eating, and another state identifies when the same individual is working or studying. Following the identification of these states, a flow diagram is developed representing a predictable sequence. Additionally, the authors illustrate the idea of clustering state sequences. The flow diagrams described in the paper identify two types of nodes: (1) source nodes and (2) sink nodes. The authors progress with the idea of generating the shortest possible path to traverse from one node to another. Based on these ideas, the authors present several theorems and lemmas. The delineated science is followed by illustrations of testing the proposed science and its results.

Overall, this is a very good paper that advances the science of algorithms used for sensors. The details of the method described are quite overwhelming. I would guess that the described science could be further tested on several use cases.