Cultural Processes

Prediction and Simulation of the diffusion of cultural processes

Human activities in each territory spontaneously form a complex network of interconnections, whether they are thought a priori consciously, or whether they are verified after the event as a result of local choices. This happens because every human activity in time tends to organize itself: anthropology avoids randomness.

It is for this reason that the result of every human activity in every moment of time draws a complex network of relationships between cultural, technological, economic, and social aspects and values. These aspects are not understandable with sectorial analyses. As in a biological cell each micro component interacts with the others and meaningfully with the global functioning of the cell itself, in the same way human activities and their results contribute to form a global design which, if an analytical study approximates significantly, will allow the analysts themselves to predict the next developments.

A significant forecasting capacity is fundamental to allow the human community to plan in a sensible way their near future: if I do not foresee well what is going to happen, any future plan I prepare will have unpredictable and often unwanted effects.

Therefore, for every community, simple or complex, to foresee is not an intellectual luxury, but an ecological necessity to survive and develop in a sustainable way.

The natural processes with which each community is confronted and the cultural processes that, unconsciously or consciously, each community determines can be represented by data sets. Data is a discreet representation of the processes within which we operate and for which we operate. Each community must collect and organize data that represent the flow of its history. Without proper data collection and organization and processing, no community will be able to plan its future in a correct and responsible way.

The most urgent task of every human community, especially if it is particularly developed, is to equip itself with instruments for:

  1. To collect and organize the data of its activities and natural processes within which these activities are carried out;
  2. Elaborate these large databases with powerful and intelligent mathematical tools to understand the real meaning of what is happening in their community and predict what, given the current situation, once understood in a subtle and profound way, could happen on different fronts;
  3. Realize a Dynamic Simulator of Scenarios (Dynamc Scenarios Simulator -DSS) able to show “in theory” the complex consequences of the different decisions that a community could take to guide its immediate future.

The tools

Complex mathematical instruments capable of performing the following operations on the data are required to perform a Dynamic Scenario Simulator (DSS):

  1. Spatial Analytics: Analyze the implicit semantics of geo-referencing of collected data. Many of the data collected to represent a process and/or structure (framework), in fact, are characterized by geographical coordinates that localize them in the territory. The statistical distribution of the latter must be analysed using appropriate methods. The methodology chosen to carry out such analyses is the Topological Weighted Centroid (TWC), a particular adaptive algorithm that can be included in the Geographic Profiling field, but that presents a very different mathematics of the classical Geographic Profiling tools and that also considers the information concerning the semantic distance of the places that analyzes and not only their physical distance.
  2. Explorative Analytics: Analyze the complex non-linear relationships between variables (attributes) and records (observations) of the collected data. To this end, specific types of Unsupervised Deep Neural Networks and specific Evolutionary Algorithms have been selected: Auto Contractive Map, New Recirculation Neural Networks, Self Organizing Map, Auto Associative Back Propagation, Evolutionary Algorithms, etc..
  3. Networks and Graph Analytics: Generate weighted, direct and a-directed graphs that selectively highlight similarities between variables and records of previously analyzed data. Minimum Spanning Tree, Maximally Regular Graph, Graph In-Out, Hierarchical Clustering and other filters have been chosen to select the significant interconnections identified by the trained matrices of the Neural Networks used in previous analyses.
  4. Simulation Analytics: To simulate situations and hypothetical scenarios on the collected data in order to analyze the possible effects of different decision making choices. Specific Neural Networks (Dynamic Associative Memories) have been chosen for these operations: Activation & Competition Systems, Spin Net, etc..
  5. Prescriptive Analytics: Be able to select targets from the database in order to apply adaptive algorithms able to calculate which variables and events should be activated and in what amount, so that the objectives set are optimized in a short time. Hybrid systems formed by Neural Networks and Evolutionary Algorithms have been chosen for these operations.
  6. Transactional Analytics: Found data with different variables and records in a single data set so that it can be analyzed comprehensively using the tools described above. The recent Theory of Impossible Worlds (TIW) has been chosen for this type of operation.