Data envelopment analysis(DEA) and regression analysis(RA) are the methods commonly used in dealing with multi-attribute decision making(MAPM). The evaluation criteria of DEA and RA are “good” and “average image”, respectively. Combining these two methods to form a unified evaluation method, namely DEARA method, can realize continuous evaluation from the “good” to “average image”. In addition, a cone ratio DEA model(C$^2 $WH) can use the cone ratio to reflect the decision maker’s preference for the sample index and decision-making unit. Based on this model, a DEARA model with a partial cone is proposed. The proposed model is applied to the analysis of the efficiency of railway transportation enterprises. The rationality of the proposed model in the multi-attribute decision-making of objective data and supervised factors is verified by examples.