Machine learning methods have been used in the Oil and Gas industry for about thirty years. Applications range from interpretations of geophysical, well and seismic responses, identification of minerals, analysis of rock samples and cores, fluid properties characterization, formation damage control, risk analysis, to well control (Alegre, 1991). In my thesis, I apply various machine learning methods for generating three well logs in shale formations, namely Nuclear Magnetic Resonance (NMR) T2 log, Dielectric Dispersion (DD) logs, and sonic travel time logs. NMR log acquired in geological formations contains information related to fluid-filled pore volume, fluid phase distribution, and fluid mobility. Raw NMR responses of the formation are inverted to generate the NMR T2 distribution responses in the geological formation, which is further processed to compute the effective porosity, permeability, bound fluid volume, and irreducible saturation of the formation under investigation. I developed two neural-network models that process conventional, easy-to-acquire logs to generate the in-situ NMR T2 distribution along 300-feet depth interval of a shale reservoir in Bakken Petroleum System (BPS). Following that, we generated DD logs. DD logs acquired in subsurface geological formations generally comprise conductivity (σ) and relative permittivity (ε_r) measurements at 4 discrete frequencies in the range of 10 MHz to 1 GHz. Acquisition of DD logs in subsurface formation is operationally challenging and requires hard-to-deploy infrastructure. I developed three supervised neural-network-based predictive methods to process conventional, easy-to-acquire subsurface logs for generating the 8 DD logs acquired at 4 frequencies. These predictive methods will improve reservoir characterization in the absence of DD logging tool. The predictive methods are tested in three wells intersecting organic-rich shale formations of Permian Basin (PB) and Bakken Shale (BS). Finally, we generated compressional and shear travel time logs (DTC and DTS, respectively) acquired using sonic logging tools. DTC and DTS logs are used to estimate connected porosity, bulk modulus, shear modulus, Young’s modulus, Poisson’s ratio, brittleness coefficient, and Biot’s constant for purposes of geomechanical characterization. Six shallow learning models, namely Ordinary Least Squares (OLS), Partial Least Squares (PLS), Least Absolute Shrinkage and Selection Operator (LASSO), ElasticNet, Multivariate Adaptive Regression Splines (MARS) and Artificial Neural Network (ANN) models, suitable for function approximation problems, are trained and tested to predict DTC and DTS logs. 8481 observations along 4240-feet depth interval of a shale reservoir in Permian Basin (PB) are available for the proposed data-driven application. ANN model performs the best among the six models. Generation of NMR T2 is the computationally most challenging and we had the least amount for data from 220-feet depth interval that made the task even more challenging; nonetheless, we obtained prediction performance of 0.85 in terms of R2. On the other hand, the generation of dielectric permittivity and conductivity dispersion logs was slightly lower in terms of computational cost as compared to NMR T2 generation, we had data from 2200-feet depth interval, and prediction performance for this log generation task was 0.79 in terms of R2 in average. Generation of DTC and DTS logs is computationally easiest among the three tasks, we had data from 4240-feet depth interval, and the prediction performance was 0.86 in terms of R2 in average.