Neurological disorders such as Alzheimer's disease and Parkinson's disease possess complex pathologies that are only partially understood. As more comprehensive and sophisticated studies are implemented in an effort to further understand the underlying pathologies of such disorders, the generation of larger and more complex quantitative output is becoming increasingly more commonplace. Extracting relevant biological insight from such output can be challenging and often requires the application of sophisticated computational tools that are capable of reducing complexity so that potentially biologically relevant patterns can emerge. In this thesis, two machine learning classification algorithms, Linear Discriminant Analysis and Random Forests(TM), are applied to a complex proteomics data set derived from a multi-subject study of human Cerebral Spinal Fluid (CSF) using a 2d-gel electrophoresis in an effort to identify novel Alzheimer's disease and Parkinson's disease biomarkers and the results are reported. Additionally, a review of recent proteomic studies focused on the discovery of novel Alzheimer's disease biomarkers within CSF is presented. Described also is a novel visualization tool iTRAQPak, which was successfully applied to the analysis of CSF based iTRAQ(TM) protein expression data sets obtained from a cohort of Alzheimer's disease subjects participating in a Phase I drug trial.