Harnessing spatio-temporal patterns in data for nominal attribute imputation

Missing data in Volunteered Geographic Information (VGI) are an unavoidable consequence of data collection by non-experts, guided by only vague and informal mapping guidelines. While various Missing Value Imputation (MVI) techniques have been proposed as data cleansing strategies, they have primarily targeted numerical data attributes in non-spatial databases. There remains a significant gap in methods for imputing nominal attribute values (e.g., Street Name) in map databases. Here, we present an imputation algorithm called the Membership Imputation Algorithm (MIA), targeting spatial databases and enabling imputation of nominal values in spatially referenced records. By targeting membership classes of spatial objects, MIA harnesses spatio-temporal characteristics of data and proposes efficient heuristics to impute the class name (i.e., a membership). Experimental results show that the proposed algorithm is able to impute the membership with high levels of accuracy (over 94%) when assigning Street Name(s), across highly diverse regional contexts. MIA is effective in challenging spatial contexts such as street intersections. Our research serves as a first step in highlighting the effectiveness of spatio-temporal measures as a key driver for nominal imputation techniques.