The BoneXpert technique recreates, from hand radiographs, the 15 bones borders consequently and afterward figures ldquointrinsicrdquo bone ages for 13 bones each. It changes the characteristic bone ages into Tanner Whitehouse or Greulich Pyle bone age. The reconstruction bone strategy consequently rejects images with anomalous bone morphology or exceptionally poor quality of image. From the methodological perspective, BoneXpert contains the accompanying advancements: 1) a generative bone reconstruction model; 2) the bone age prediction from surface score, shape, and force gotten from central segment examination; 3) the concept of consensus bone age which characterizes each bone’s age as the bone age best gauge of alternate bones in the hand; 4) a typical female and male bone age model; and 5) the unified GP and TW bone age modelling. BoneXpert is produced on 1559 images. Examined for foreseeing a compound's quantitative or unmitigated natural action in view of a quantitative portrayal of the compound's atomic structure. Random Forest is a troupe of unpruned characterization or relapse trees made by utilizing bootstrap tests of the preparation information and random selection of feature in induction tree. Prediction is made by collecting (larger part vote or averaging) the ensembles forecasts. We assembled prediction models for six cheminformatics informational sets. Our examination exhibits that Random Forest is a capable apparatus equipped for conveying execution that is among the most precise strategies to date. We likewise introduce three extra elements of Random Forest:  worked in execution appraisal, a measure of relative significance of descriptors, and a measure of compound comparability that is weighted by the relative significance of descriptors In this thesis use the different group images and extract the geometric features and then use principle component analysis use extracted features and class in classification method and compare the results in our approach SVM with RBF kernel play important role in classification of images and predicting of images. In this done experiment SVM with RBF show 78% accuracy which has significance different from other methodology like linear regression and voted regression.