Diabetes has since become global pandemic – which must be diagnosed early enough if the patients are to survive a while longer. Traditional means of detection has its limitations and defects. The adoption of data mining tools and adaptation of machine intelligence is to yield an approach of predictive diagnosis that offers solution to task, which traditional means do not proffer low-cost-effective results. The significance thus, is to investigate data feats rippled with ambiguities and noise as well as simulate model tractability in order to yield a low-cost and robust solution. Thus, we explore a deep learning ensemble for detection of diabetes as a decision support. Model achieved a 95-percent accuracy, with a sensitivity of 0.98. It also agrees with other studies that age, obesity, environ-conditions and family relation to the first/second degrees are critical factors to be watched for type-I and type-II management. While, mothers with/without previous case of gestational diabetes is confirmed if there is: (a) history of babies with weight above 4.5kg at birth, (b) resistant to insulin showing polycystic ovary syndrome, and (c) have abnormal tolerance to insulin.

A major challenge today in communication and over various communications medium is the wanton havoc wreaked by attackers as they continue to eavesdrop and intrude. Young and inexperienced academia are today faced with the challenge of journal houses to send cum have their articles published. The negative impact thus, of predatory and hijacked journals cannot be over-emphasized as adversaries use carefully crafted, social engineering (phishing attack) skills – to exploit unsuspecting and inexperienced academia usually for personals gains. These attacks re-direct victims to fake pages. The significance of the study is to advance a standard scheme/techniques employed by phished (predatory/hijacked) journals to scam young academia and inexperienced researchers in their quest for visibility in highly impactful indexed journals. Thus, our study advances a decision-tree algorithm that educates users by showing various indicators cum techniques advanced by predatory and hijacked journals. We explore journal phishing attacks employed by such journals, targeted at young academia to adequately differentiate also using web-page ranking. Results show the classification algorithm can effectively detect 95-percent accuracy of journal phishing based on journal metric indicators and website ranks.