Estimating GDP in absence of historical GDP data using SMT solvers and machine learning clustering algorithms
Gross Domestic Product (GDP) is a measure of interest that reflects the values produced by an economy in a given year. Though it is crucial in policy formulation, business decision-making, etc., GDP data is unavailable in some areas. This project presents a solution to estimate the GDP in such a scenario. In particular, we form an equation containing a set of constraints that should be satisfied. The details of this step are beyond the scope of this summary. In the next step, we solve the equation using an SMT-solver for a prespecified time. Afterwards, we cluster the solutions found by the SMT-solver and suggest the centroids to the expert. As the number of centroids is small, the expert examines the potential solutions and chooses the best estimation.
Application Language(s): N/A
Programming Languages and Technologies: Python, Z3, scikit-learn
Member(s): Parham Hasaninia, Taha Rostami, Towhid Firoozan Sarnaghi