Geo-mechanical parameters are very important in petroleum industry. In order to obtain the geomechanical parameters, the sonic log (compressional and shear velocities) should be available. In many cases, the sonic log is not available or missing from the log data, for that cases the existing correlations are used to predict sonic time, most of the existing correlations use the compressional velocity to predict the shear velocity.
The objective of this paper is to develop simple and accurate mathematical model to determine the compressional and shear sonic times using log data (gamma ray, bulk density, and neutron porosity). These three logs are commonly conducted at every well and they are always available. Three artificial intelligence techniques namely; ANNs (Artificial Neural Networks), ANFIS (Adaptive Neuro Fuzzy Inference System), and SVM (Support Vector Machines) are used. Finally, an attempt has also been made to converge the results into one simple empirical correlation using the weights of ANN model in order to make a generalized model that can be used for field applications.
The results obtained showed that ANNs model successfully predict the compressional and shear sonic times from log data with 99% accuracy giving correlation coefficient of 0.99 when compared to actual field data.
Accurate prediction of rock mechanical parameters is critical for alleviating the risk associated with the drilling and maximizing the well productivity. Rock Mechanical parameters are used in optimizing the well placement, wellbore instability evaluation Moos et al., 2003 , completion design, draw-down limits to avoid sand production Santarelli et al. 1989 , hydraulic fracturing, and many other applications. Inaccurate determination of rock mechanical parameters leads to heavy investment on decisions and field development, Abdulraheem et al. 2009 .
Generally elastic parameters can be estimated from the sonic logs (compressional and shear velocities) using regression dependent correlations, Potter and Foltinek 1997  and Pickett 1963 . Log derived, dynamic rock parameters should be calibrated to core derived static (laboratory) parameters, because the static measurements more accurately represent the in-situ reservoir mechanical properties, Edlmann et al. 1997 . Many of the correlations for elastic parameters determination were been reviewed by Chang et al. 2006 . They concluded that calibration of the empirical correlations between physical parameters and strength is necessary for any correlation to be used with degree of confidence. Further, shear and compressional wave velocities data is not always available from well logs, making the problem more difficult.