My Journey
Brain Tumor classification and detection from MRI images using CNN based on ResUNet Architecture.
Jacobs University Bremen. May 18, 2020
Bremen, Germany
e Convolution Neural Network (CNN) deep learning algorithm based on ResUNet architecture is proposed for detecting the tumor and marking the area of their occurrence. There are several advantages to using these proposed architectures for segmentation tasks. First, a residual unit helps when training deep architectures. Second, feature accumulation with recurrent residual convolutional layers ensures better feature representation for segmentation tasks.The automatic brain tumor classification is very challenging task in large spatial and structural variability of surrounding region of brain tumor. The method proposed accuracy 96% on the test data.
Big Data Challenge 2020.
Jacobs University Bremen. May 18, 2020
Bremen, Germany
The project is a machine learning solution for the supervised classification task provided by the University of Bremen as The Bremen Big Data Challenge. The task was to predict the appropriate classification model for a given data for human activity recognition with a high degree of accuracy. The available data set is made up of signals detected by motion sensors that can be placed on the human subjects' leg.The optimization method is applied to the data set. Due to the high features of the data set, the feature extraction method - Principal Component Analysis is performed. After the extraction, K Nearest Neighbour, Random Forest Approach and MLP Classifier methods are used to obtain test results of 80%, 78% and 75% accuracy using Scikit Learn in Python.
Design and Analysis of Lotka Volterra Model in 2 and 3 Dimensions.
Jacobs University Bremen. May 18, 2020
Bremen, Germany
This project is an introduction to a non-linear dynamical systems. Here, I desinged a mathematical model for predator-prey model in two dimensions using ODE, Matplotlib in Python. Analysed the behaviour of the model by changing the parameters such as growth rate, death rate by observing direction field and trajectories in the phase plane using Matplotlib. Furthurmore, I designed predator-prey model for 3 dimensions using ODE and analyses the behaviour of the direction field and trajectories in a phase place by changing the parameters.
Modelling and design of non- linear naturally occurring phenomena -Synchronization of Fireflies illumination and response to external stimulus.
Jacobs University Bremen. May 18, 2020
Bremen, Germany
I designed a mathematical model for Firefly in the presense of a stimuli using Ordinary Differential Equations in python. I analysed the change in the direction field and trajectoies near the fixed point of the model in the presence of one stimulus. Lastly, I designed the mathematical model for a Firefly in response to two stimuli and performed fixed point analysis by changing the parameter using Matplotlib in python.
Sentiment Analysis of movie reviews taken from RottenTomatoes.com which involved multi-class data classification to predict the sentiment of the Phrases from the sentences of the movie reviews given by the customers, on an arbitrary sentiment scale of 0 – 4.
Jacobs University Bremen. June 14, 2020
Bremen, Germany
I predicted sentiment of the reviews by using logistic regression, decision tree, and random forest obtaining an accuracy of 83%, 99%, 98% accuracy respectively. For the text processing, words which were common in context in the corpus were located close to one another in a vector space. For example “Have a nice day.” and “Have a great day.” Here great and good was placed closer in the vector space because they convey similar meaning in this context.
DIABOT-An Expert System in the management of Diabetics using Artificial Intelligence."
International Journal of Computer Applications (0975 – 8887) Volume182 – No.4, July 2018.
Banglore, India
Expert Engine for Diabetics using Artificial Intelligence and hardware to measure twelve parameters (like blood pressures, heart rate, etc) using sensors.The inference engine used these data types to compare it with health metrics present in the knowledge base using backtracking and searching algorithms obtaining accuracy of 83% using python and prolog.
