A new approach to building the interindustry input-output table using block estimation techniques

摘要: We present a new approach to estimating the interdependence of industries in an economy by applying data science solutions. By exploiting interfirm buyer-seller network data, we show that problem is similar uncovering latent block structure literature. To estimate underlying with greater accuracy, propose extension sparse model incorporates node textual information and unbounded number interactions among them. The latter task accomplished extending well-known Chinese restaurant process two dimensions. Inference based on collapsed Gibbs sampling, evaluated both synthetic real-world datasets. proposed improves predictive accuracy successfully provides satisfactory solution motivated problem. also discuss issues affect future performance this approach.