Coronavirus has spread all over the world and brought significant impact to everyday life. This highly infectious virus calls for new strategies and policies that are able to suppress its transmission rate. One of the ways is by lockdown. However, lockdown is not an infallible solution and it brings with it negative impacts on society in terms of socioeconomic and mental health aspects. In addition, several factors should also be taken into account before the implementation such as demographic conditions, health services and COVID-19 case data itself. In general, a lockdown decision can be determined by looking at the level of risk of an area. Naturally, this assessment of levels of risk must be based on parameters affecting the lockdown. In this paper, we propose a multicriteria recommender system model that can calculate the value and level of risk for each region by taking into account several parameters from different databases. This model is also equipped with the process of weighting using the Analytical Network Process (ANP) method to determine interdependence and feedback between parameters. From the experiment conducted on 27 cities and districts in West Java, it was found that 15% of the regions were in the high-risk category, 41% of the regions have medium risk and 44% of the regions have a low risk.