Analysis of the human wayfinding problem in word networks

OData support
Dr. Gulyás András
Department of Telecommunications and Media Informatics

The networks are becoming more involved in our everyday life; therefore, it is increas-ingly important for us to understand their function. We can encounter such networks in each discipline, e.g.: sociology, technology and biology. We can find them in our daily lives, whether it is a network of airports or the Internet where the goal is to serve the user smoothly. This is achieved by reducing the load on each unit, i.e. the amount of traffic corresponding to their capacity, thus avoiding any failure. Cancer researchers are seeking for a similar method in which individual cells destroy cancer by being overload-ed due to cascading failure. To control the traffic or load, there is a need for a kind of routing strategy that can regulate the network load. This routing strategy can determine how the load (a packet) travels between the initial and end points.

I’ve studied the human routing strategies in a predetermined network of words with a word-morph game. The data was provided by users of a game developed specially for this purpose. The game has given the equally long words. From the initial word, you must get to the target word as few steps as possible by changing only a single letter at each step. With the agreement of the players, I have collected anonymously information about the gaming habits: game number, time spent, solutions, etc.

In my research, it turns out that the strategy which we decide what route we want to travel to solve the problem is based on regularity. To be able to get to the destination, first we navigate to a so-called landmark (an intermediate station known to us). All the landmarks have high centrality, which means they are important elements of the net-work.


Please sign in to download the files of this thesis.