Blast-produced ground vibrations is one of the fundamental problems in mining industry as well as construction operations. Therefore, the prediction of ground vibration components is very important in the minimization of blast related complaints. This paper presents the results of simple and multiple linear regression analyses for the comparison of the most widely used peak particle velocity prediction equations. For this purpose, the results of blast-induced ground vibration measurements by different authors with the measurements conducted within this study were taken into account together. In the statistical analyses, a total of 247 regression models were analysed by using a statistical software. Then, the site-specific constants were determined by the regression analyses of scaled distance and peak particle velocity data pairs according to different approaches and considered cases were evaluated statistically. By evaluating the statistical results, a new prediction equation based on scaled distance concepts was proposed. In addition, new attenuation formulas were proposed for the determination of the site-specific constants. The analyses have shown that if scaled distance concept itself is used for inelastic attenuation factor to predict PPV, the highest values of coefficient of determination are obtained.