|
Hi 👋, I am Vignesh Muthukumar, pursuing my Masters in Computer Science 🎓 from the North Carolina State University. I am passionate about developing scalable Software applications blending novel AI methodologies to create systems that are socially impactive. My academic research focuses on Educational AI and Intelligent Tutoring Systems, studying the behavioral patters of student's engagement with academic courses. I have about 2.5 years of industry experience working as a Software Engineer. I had interned with Salesforce in Summer 2022 working with the Commerce Cloud Organization. Prior to that, I worked with an e-commerce based startup called Ninjacart for over 2 years, in their SCM-Tech team creating sustainable Last Mile Logistics (LML) products. With a vision to create impact through cutting edge technology, I am actively looking for Software Engineer opportunities. Away from work, I like to hike ⛰ and bike 🚲. Recently picked up the habit of reading books 📚 and playing piano 🎹. Hit me up to discuss interesting things!
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Paper /
Abstract /
Slides /
Code
Massive Open Online Courses (MOOCs) has seen a dramatic increase of participants in the last few years with an exponential growth of internet users all around the world. MOOC allows users to attend lectures of top professors from world class universities. Despite flexible accessibility, the common trend observed in each course is that the number of active participants appears to decrease exponentially as the week’s progress. The structure and nature of the courses affects the number of active participants directly. A comprehensive review of the available literature shows that very little intensive work was done using the pattern of user interaction with courses in the field of MOOC data analysis. In this paper, we take an initial step to use the deep learning algorithm to construct the dropout prediction model and produce the predicted individual student dropout probability. Additional improvements are made to optimize the performance of the dropout prediction model and provide the course providers with appropriate interventions based on a temporal prediction mechanism. Our Exploratory Data Analysis demonstrates that there is a strong correlation between click stream actions and successful learner outcomes. Among other features, the deep learning algorithm takes the weekly history of student data into account and thus is able to notice changes in student behaviour over time. |
![[NEW]](images/new.png)
|
|
|
The North Carolina Environment Climate Observing Network is a sensor network that collects data from 43 weather stations to assess 23 weather measure values. An automatic Quality assurance (QA) system flags the measurements in the dataset, and they are manually inspected for mistakes. The QC dataset is imbalanced, with low erroneous data and a majority of data that has been correctly flagged. This study presents with a novel approach of incorporating season-specific data and resampling the provided dataset to classify the erroneous data points. A robust classification based algorithm is experimented to work on the time series based imbalanced dataset. There is a strong dependency of how different weather measure changes with time. This poses a novel challenge of wrapping the measure values across each weather station over time periods to generate a predictive model. Since the distribution of erroneous points are very less, they pose a challenge of being considered as an outlier which is handled effectively in this study to correctly classify such imbalanced dataset. |
|
The goal of the study is to utilize inertial measurement units (IMU) acquired from the lower leg through sensors, to distinguish different types of terrain. The units have accelerometer and gyroscope readings gathered across the x,y, and z axes from a system linked to a prosthetic limb, and the terrains are widely labeled by specialists as follows: (0) - indicates standing or walking in solid ground, (1) - indicates going down the stairs, (2) - indicates going up the stairs, (3) - indicates walking on grass. By extracting some patterns in their motion, the data is rolled into smaller time windows and fed into temporal based neural network architectures for predicting the different types of terrains. |
|
|
|
|
|
|
|
|
|
|
|
|