The dataset for the joint shear strength of beam-column joints subject to cyclic loading.

The learning dataset on which the prediction model is trained is the most critical factor in the success of the proposed framework prediction models. Experiments with data on joint shear response, including joint shear strength and distortions, were chosen for the database to guarantee that the specimens included in the joint shear strength prediction had acceptable amounts of shear deformations. Subassemblies with wide beams, slabs, and/or transverse beams are also included in the database. Furthermore, specimens with eccentricity (e) between the centrelines of the longitudinal beams and the column are included in the database to explore the influence of eccentricity on joint shear strength. Also, the axial load acting on the column (N), column dimensions (bc, hc), beam dimensions (bb, hb) concrete cylinder compressive strength (fc`), yield of beam longitudinal reinforcement (fy) and beam longitudinal reinforcement ratio (ρ) are factors affecting the joint shear strength. All joint subassemblies in this database have strong column-weak beam behavior. There are 98 specimens in the experimental database from 18 different research projects. Table 1 shows the database of studied beam-column connections. The concrete compressive strength (fc`) and volumetric joint reinforcement ratio (ρcore), which reflects the confinement supplied by the transverse reinforcement, are the most relevant characteristics.