A passionate Machine Learning enthusiast aspiring to find solutions for business & community challenges
LinkedIn
Convolutional Neural Network uses the eye-tracking fixmap and heatmap data of the normal and ASD class groups. Tuned the model for best hyperparameters and improved the AUC ROC value by 20% from the baseline value
EDA | DATA PRE-PROCESSING | MODEL BUILDING | MODEL TRAINING | HYPER-PARAMETER TUNING | PREDICTION | MODEL EVALUATION |
View on Github
Determine the selling price of a house taking into account of many factors like location, size, age of house etc,. Extensive Exploratory Data Analysis on this Kaggle dataset revealed some of the misinterpreted/low variance variables and aided in the process of Feature selection for model training
EDA | DATA CLEANING | FEATURE ENGINEERING | DATA PRE-PROCESSING | MODEL TRAINING | GRADIENT BOOSTING | HYPER-PARAMETER TUNING | MODEL EVALUATION |
View on Github
Predicting Autism Spectrum Disorder(ASD)in children using face images of children. Trasfer Learning method is incorporated in the Convolutional Neural Network model to efficiently extract features by connecting the pre-trained VGG16 to the fully connected layer
EDA | DATA PRE-PROCESSING | MODEL BUILDING | TRANSFER LEARNING | MODEL TRAINING | PREDICTION | MODEL EVALUATION |
Page template forked from evanca