Lare Adeola

This paper presents a novel hybrid deep learning model that combines Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs) for accurate age and gender classification from iris images. Using the GMBAMU-IRIS dataset, the model leverages CNNs for spatial feature extraction and RNNs for sequential pattern recognition. The proposed model achieved an age classification accuracy of 85.68% and a gender classification accuracy of 98.95%, outperforming traditional methods. These findings highlight the potential of hybrid models in enhancing biometric recognition systems. 1 Introduction Biometric recognition systems have emerged as critical components in various security and authentication applications, providing reliable and efficient means of identifying individuals based on their physiological and behavioral traits. Among the array of biometric modalities, the human iris stands out due to its unique patterns and remarkable stability throughout an individual's life. Iris recognition systems leverage the distinct textural patterns within the iris to achieve high levels of accuracy and reliability in personal identification and verification tasks [1][2]. Beyond the primary goal of identification, there is a growing interest in using iris images for soft biometric traits, such as age and gender classification. Accurate age and gender classification can significantly enhance the functionality of biometric systems by providing additional contextual information [3]. This additional layer of data can improve user experience , enable demographic-specific applications, and enhance the robust-ness of security systems by adding another factor of authentication [4].

Suicidal Detection and treatment from the clinical and public health perspective is reactive. For an action whose consequences are irreversible, a reactive approach to the problem cannot be the answer. A proactive approach is needed to solve and detect Suicidal Intent. Social Media has become the television and diary of millennials and gen-z alike, hence it's imperative to create techniques and approaches to study their actions in this particular space. This research involved creating Document Similarity Algorithms from Corpora mined from the Twitter Developer API. Making the data unique to this platform; a methodology design involving validating data at various spectrum and selecting an appropriate threshold to classify the similarity levels were created as well as a lexicon unique to the Twitter Dataset. With an accuracy score of 84%, the Jaccard Document Similarity Algorithm was able to spot Suicidal intent from User's Tweets and with an accuracy of 93% was also able to spot non-suicidal intent. The Jaccard model seemed to be the most durable and computationally efficient for the problem and was chosen as the algorithm for detecting Suicidal Tendencies in Users' Tweets.

Understanding the sentiment nature of Trade Policy
documents is a crucial step in both fostering global and regional
trade. Moreso, the sentiments present in trade policy documents
offer valuable insights into the direction, attitudes, and political
atmosphere amongst member countries. This research aims to
provide a sentiment analysis of one of Africa’s crucial trading
blocs, COMESA, using Natural Language Processing. By using
Machine Learning techniques, we arrived at the conclusion
that the COMESA Trade Policy document has mostly positive
sentiments with preferences for ’Member States’, ’Transit of
Goods’ and ’Common Market’ to name a few. The research
will lay the groundwork for future sentiment analysis of other
global and regional trading blocs.

Maternal health has become an increasing health problem, especially in the Global South, where the risks, although reducing but not fast enough, continue to register high maternal mortality rates. This research work sought to provide an ergonomic and easily accessible solution by creating a machine learning web application that will be able to detect maternal health with an accuracy of 84%. The web application is a simple form where the user has a set list of questions and answers, based on those choices, a prediction is made. The research hopes this tool will be deployed in areas with high maternal mortality rates in order to bring down the avoidable risks of deaths experienced by pregnant women.

Various approaches exist when building a detection model to capture Cyber-Threats but most of this approaches employ a post-active methodology-trying to detect the threats after they have occurred. We aimed to develop a model that would employ a pro-active approach by understanding the semantic and linguistic nature of their source of origin-urls and from there building a classifier that can identify potential threats. Our Decision Tree classifier achieved an accuracy of 95% on the test set showing its potential to detect cyber-threats in real life scenarios. And since our model uses a classical algorithm as opposed to deep learning methods, our model would be computationally less expensive and lightweight making it easy to deploy in real world web applications.

Despite the preponderance of sentiment corpora in different domains in technology, there remains a gap in finding such sentiment data in the domain of Augmented/Virtual Reality. In this paper, we created a Corpora of Sentiments based on AR/VR Headsets and in the process we were able to extract specific pain-points for AR/VR enthusiasts and users. We believe that this corpora and it's subsequent corpus can be used for further sentiment analysis which would aid product development and innovation in the realm of AR/VR especially as it relates to AR/VR headsets.

’HOMESWEET’ is a deep learning model created with the YOLO object detection algorithm that has been trained to detect certain human physical states that could result in road accidents and deaths. The model achieved an accuracy score of 83% and a Precision rate of over 90% but had a fairly modest Recall rate of just over 70%.
The model can be deployed in various other applications as it was able to detect not only facial cues but other micro-expressions and gesticulations that lead to the various states; in particular, in this research, it was created to detect fatigue, drowsiness and lack of total concentration while driving.

Numerous studies and research work has been undertaken in the area of creating predictive models for studying Bank Churn. In these studies, the end goal was to create a high accuracy predictive model; while this is commendable, this research focuses on creating an architecture for a predictive model by aggregating the power of various predictive models. The architecture and model proposed in this paper achieved an accuracy of 91% in the test data (35% of the original data set), and an AUC of 96% - confirming the generalized nature of the model. Also, various feature extrapolation techniques were introduced which provide valuable insights to the banking sector.

There has been a significant revolution seen by AI getting incorporated into the management and customer relations of companies. The research of the present Artificial Intelligence (AI) Revolution that influences a variety of fields i.e. video games are the topic of the article. AI systems such as machine learning and data analytics can help brands understand consumer behaviour in much greater detail; hence, companies can better reach and interest potential consumers through personalized marketing plans and campaigns. What is more, this is another case of citizen science projects that can host a large number of artisanal anglers who can together provide data that can make the research wider-reaching. This is when the conclusion is reached, which means, for gaming neither marketing nor game-play is the energy source. The proposed scheme improves the level of customer accuracy and tackles trends timely as well as creates slight space for real-time communication by applying neighbour-based recommendation techniques, neural networks, and sentiment analysis. Its supremacy over the conventional methods of statistical significance is highlighted through the advent of predictive analytics and dynamic pricing approaches. The advantage of deploying natural language processing (NLP) is that it helps to understand what the customers mean with how they write. Measuring the key performance indicators at the end of this approach can be called the method of adaptation and flexibility which makes digital marketing not just refer only the success but also turn to the happiness of customers.