Automated opinion of Alzheimer’s complaint (announcement) using audio signals has surfaced as a promising exploration area because of its non-invasiveness and cost effectiveness using the openSMILE toolkit to extract GeMAPSv01b features from audio recordings of cases with AD announcement and non-AD. The uprooted features are also used to train the different machine learning models. This work focuses on training Logistic Regression, SVM and Random Forest to classify cases as AD announcement or non-AD. The experimental results show that logistic regression achieved the accuracy of 76.47 in classifying cases, followed by SVM and random forest with rigor of 70.58 and 58.82, independently. This study suggests that audio- grounded automated opinion of announcement using machine learning algorithms has the implicit ability to give an effective non-invasive and cost-effective webbing system for early finding of announcement.