Depression is a global disorder with serious consequences. With more depression-related data and improved machine learning, it may be possible to build intelligent systems that can detect depression early on. This research uses the burns depression checklist as the gold standard for diagnosing depression and the support vector machine, decision tree, and light gradient boosting method as algorithms to create models capable of diagnosing depression on a data-set of 604 surveyed participants. This research demonstrates the efficiency of machine learning algorithms within the field of mental health. This paper serves to increase the body of knowledge by training insufficiently researched algorithms on a commonly used depression detection data-set with the goal of reaching or surpassing the level of performance seen in current research. This experimental research has found the decision tree classifier to be the best approach for predicting depression with an accuracy of 95.66% while that of the support vector machine classifier and the light gradient boosting classifier are 91.48% and 94.58%, respectively. The techniques presented in this paper perform better than those being used in current machine learning research. This research study may support the clinicians in determining what attributes are most crucial in diagnosis of depressed individuals as well as improve the health of the general populace.

The advent of convolutional neural networks (CNNs) to the development of face recognition system has been a game changer in the field of computer vision and pattern recognition. This research work uses a pre trained MobileNet-V1 model to develop an effective CNN model capable of high performance. We also tackle several common facial recognition challenges which include occlusions, illumination variations, make-ups, pose variation and ageing through the use of several improvement techniques. The techniques include adopting a less computationally costly approach, transfer learning and hyper-parameter finetuning. The Top-1 accuracy 70.6% and Top-5 accuracy 89.5% of the base MobileNet-V1 model has been improved using these techniques to achieve training accuracy of 95% and accuracies of 96.4%, 98.0% and 99.1% on the Pins face recognition dataset, FaceScrub data-set and LFW data-set, respectively. The work done so far illustrates the need for further research into improvement techniques for convolutional neural networks.