Jaison Menezes

There are many problems that are present in existing applications. Most of the applications will have one or the other feature missing in it, hence an application is required to cater to all the needs of a user. Human fitness tracking and activity recognition basically is an android application that can be used by millennials to keep a track of their health and fitness. The users are just required to enter their details like age, height and weight. Using the details, the of users the… Show more

There are many problems that are present in existing applications. Most of the applications will have one or the other feature missing in it, hence an application is required to cater to all the needs of a user. Human fitness tracking and activity recognition basically is an android application that can be used by millennials to keep a track of their health and fitness. The users are just required to enter their details like age, height and weight. Using the details, the of users the application will calculate the BMI index of the users. The users are required to track their blood pressure and heart rate using a simple method by placing their fingers on the camera of the smartphone and using the flashlight. Simple Image processing methods are used to calculate the heart rate and blood pressure. Using this the app calculates the metabolic rate of the user. Then by using KNN algorithm the app predicts certain diet plans and workouts for the users. Also, the steps taken by the user is also tracked which also adds in the calculation of metabolic rate, which in turn gives a more accurate and specific diet plans. Show less

Teachers have a problem of entering marks manually in a website and calculate
the Course Outcome(CO) and Program Outcomes (POs) our system make this
task less tedious , easy and simple by directly uploading simple excel sheet ,
which contains all the marks of a student analysing and calculating the Course
Outcome (CO) and then mapping it with the corresponding Program Outcomes
(POs)

Context: Developed a machine learning model for the HR department of a large software company as part of their "Proactive Retention" initiative. The goal was to predict the likelihood of employee turnover using key features from past employee data, enabling HR to intervene before employees leave.
Problem Statement: Modeled the employee retention problem as a binary classification task. The objective was to predict whether an employee is likely to leave, using historical data and factors such… Show more

Context: Developed a machine learning model for the HR department of a large software company as part of their "Proactive Retention" initiative. The goal was to predict the likelihood of employee turnover using key features from past employee data, enabling HR to intervene before employees leave.
Problem Statement: Modeled the employee retention problem as a binary classification task. The objective was to predict whether an employee is likely to leave, using historical data and factors such as department, salary, tenure, satisfaction, workload, and promotion history.
Key Features:
Departmental and salary data
Number of projects, average monthly hours
Tenure and promotion history
Mutual evaluations: satisfaction scores, last performance evaluation, and formal complaints
Approach:
Data Preprocessing: Cleaned and transformed the dataset of 14,249 observations. Handled categorical variables such as department and salary through encoding, and standardized numerical features like average monthly hours and satisfaction scores.
Modeling: Utilized logistic regression as the primary model due to its interpretability. Additionally, built models using Random Forest and Gradient Boosting for comparison.
Evaluation: Focused on model interpretability and prediction probabilities, using metrics such as the confusion matrix and ROC-AUC for thorough evaluation. Achieved a high ROC-AUC score, showing strong performance in distinguishing employees who are likely to leave.
Outcome: Delivered a trained and interpretable model to HR, providing probabilities for each employee's likelihood of leaving. This enabled HR to focus proactive retention efforts on employees most at risk, optimizing interventions. Show less

● Built a CNN model to detect pneumonia from chest X-ray images, aiming to assist radiologists in early diagnosis with high accuracy.
● Preprocessed the dataset of chest X-rays by resizing images, applying data augmentation, and normalizing pixel values for effective model training.
● Developed a multi-layer CNN architecture with convolutional layers, pooling layers, and ReLU activation, followed by a fully connected layer for classification.
● Trained the model using binary… Show more

● Built a CNN model to detect pneumonia from chest X-ray images, aiming to assist radiologists in early diagnosis with high accuracy.
● Preprocessed the dataset of chest X-rays by resizing images, applying data augmentation, and normalizing pixel values for effective model training.
● Developed a multi-layer CNN architecture with convolutional layers, pooling layers, and ReLU activation, followed by a fully connected layer for classification.
● Trained the model using binary crossentropy loss and Adam optimizer, achieving high accuracy in detecting pneumonia.
● Aided early diagnosis by providing quick and accurate predictions on whether a person has pneumonia, supporting faster medical interventions. Show less