https://www.linkedin.com/in/rkshiyaniya
https://github.com/rkshiyaniya
Data Science Enthusiast with 1+ years of working experience in the field of Data Science, Machine Learning, and Deep Learning (Computer Vision & Natural Language Processing).
Key Achievements & Responsibilities :
- Analyzed and Visualized Data for better in-sight
- Performed Data Preprocessing & Data Preparing tasks
- Used XGBoost & Random-Forest classifier to train a model
- Evaluate Model Performance for generalization
- Deployed Predictive end-to-end ML Pipeline
Key Achievements & Responsibilities :
- Trained on Machine Learning & Data Science Concepts
- Applied Exploratory Data Analysis and Data Preprocessing techniques on various types of Dataset
- Worked on live industry assigned projects
- Learnt to design complete Data Science Project
- Implemented end-to-end Data Science projects from Data Preprocessing to Build Predictive Model and deployed on local server
Key Achievements & Responsibilities :
- Worked on Computer Vision based Doctor’s Handwritten Prescription Recognition Project
- Researched and Designed Project workflow
- Used CNN + LSTM based architecture to make a base model in TensorFlow
Tech. Stack :
- Python
- NumPy
- Pandas
- Matplotlib
- Seaborn
- Sci-kit Learn
Categories :
- Supervised Learning
- Classification Algorithms
- Machine Learning
- Data Science
- Exploratory Data Analysis
- Data Preprocessing (Handling Data Imbalance, Outliers)
- Data Visualization
- Feature Selection
- Machine Learning Model Building
- Model Evaluation
- Dataset used : Click here to download
- I have uploaded 2 versions of this project.
- Version 1 : This Notebook contains simple method for feature selection based on correlation with target attribute.
- Version 2 : This Notebook contains logistic regression method for feature selection based on column's accuracy.
- Able to got ~100% accuracy.
For more details and Source Code Visit my Repo : Github Link
- Dataset used : Click here to download
- Project contains in-depth insight into Dataset - Exploratory Data Analysis, Visualization and Data Preparation.
- Tried different algorithms for classification and got ~98% accuracy.
For more details and Source Code Visit my Repo : Github Link
- Dataset used :
- Train - Click here to download
- Test - Click here to download
- Project contains - Data Preprocessing by Scaling, Transforming into One-hot vectors, Data Preparation for model bulding and Model evaluation.
For more details and Source Code Visit my Repo : Github Link
- Dataset used : Click here to download
- This Project contains in-depth explaination of K-means clustering algorithm with it's working visualization on randomly generated dataset.
- Also, Used K-means to segment customers to 3 different Clusters.
For more details and Source Code Visit my Repo : Github Link
- Dataset used : Click here to download
- This Project contains end-to-end implementation of Decision Tree Classifier with printing tree also.
- Tree : Click here to view
- This Project contains data preprocessing, model building and model evaluation.
- Got ~98%+ accuracy.
For more details and Source Code Visit my Repo : Github Link
- Dataset used : Click here to download
- This Project contains in-depth Exploratory Data Analysis and Visualization about Nutritions in McDonald's Menu.
For more details and Source Code Visit my Repo : Github Link
Tech. Stack :
- Python
- TensorFlow/Keras
- NumPy
- OpenCV
- PIL (pillow)
- tkinter
- Sci-kit Learn
- Matplotlib
- DNN Caffe Models - face detection
- mobilenet_v2 base model with pre-trained weights of 'imagenet'
Categories :
- Image Classification (Computer Vision)
- Deep Learning
- Transfer Learning
- Real-time Face Detection
- Image Augmentation
- Neural Network Architucture Implementation
- Model Evaluation
- It's binary class classification task - (People Wearing Mask & Without Mask)
- For Face Detection DNN based caffe model has been used.
- For Model training I have used Transfer Learning with 'mobilenet_v2' Neural Network base model with pre-trained weights of 'imagenet'.
- Made it Real-time with the help of OpenCV.
- It's multi-class classification task - (Predict digit between 0 to 9)
- Dataset Used : MNIST digit
- Deep Learning Model has been built in TensorFlow/Keras from scratch and trained using CNNs.
- With the help of OpenCV it's possible to detect Multiple Digits in Canvas made in tkinter.
- Detected digits are passed to Model for Prediction.
- It's multi-class classification task - (Predict Rock, Paper & Scissors)
- Animated Dataset has been used.
- Able to got ~98% Validation accuracy.
- Correclty classify all the unseen images except only 1.
- Note : Data Label - Paper 0, Rock 1, Scissors 2
- It's multi-class classification task - (Predict digit between 0 to 9)
- LeNet Architecture has been used for Image Classification on MNIST handwritten digit dataset.
- It's multi-class classification task - (Predict between 10 different classes)
- MiniVGGNet Architecture has been used for Image Classification on cifar10 dataset.
For more details and Source Code Visit my Repo : Github Link
Tech. Stack :
- Python
- TensorFlow/Keras
- NumPy
- Matplotlib
Categories :
- Image Segmentation (Computer Vision)
- Deep Learning
- Neural Network Architucture Implementation
- Model Evaluation
Use Pretrained VGG-16 network for the feature extraction path, then followed by an FCN-8 network for upsampling and generating the predictions. The output will be a label map (i.e. segmentation mask) with predictions for 12 classes. Trained the model on dataset contains video frames from a moving vehicle and is a subsample of the CamVid dataset.
This notebook illustrates how to build a UNet for semantic image segmentation. Trained the model on the Oxford Pets - IIT dataset dataset. This contains pet images, their classes, segmentation masks and head region-of-interest. Detailed Explaination has been presented in the notebook itself.
For more details and Source Code Visit my Repo : Github Link
This repository contains Various Techniques that can be used for object detection.
Tech. Stack :
- Python
- TensorFlow/Keras
- object_detection API
- Matplotlib
- NumPy
- Other CV related libraries
Categories :
- Object Detection (Computer Vision)
- Deep Learning
- Image Inference
- Fine Tuning
- Eager Mode training
- Image Annotation
- TensorFlow Hub
- Detailed explaination has been presented in the respective notebook itself.
A notebook for object detection with the help of fine tuning of a (TF2 friendly) RetinaNet architecture on very few examples of a novel class after initializing from a pre-trained COCO checkpoint. Training runs in eager mode.
A notebook for "out-of-the-box" object detection model from TensorFlow Hub and inference on images.
For more details and Source Code Visit my Repo : Github Link
Tableau Profile : Click-here
For more Details : Click-here
- Programming Languages -
- Proficient : Python
- Familiar With : Java & C
- Working Knowledge of Tableau
- SQL
- Probability & Statistics
- Exploratory Data Analysis -
- NumPy
- Pandas
- Data Visualization -
- Matplotlib
- Seaborn
- CV related Library -
- OpenCV, PIL, Other Utilities
- Machine Learning -
- Sci-kit Learn
- Machine Learning Areas -
- Hands-on Experience with Classification, Regression Algorithms
- Familiar with Clustering Algorithms
- Deep Learning Framework -
- TensorFlow
- Keras
- Deep Learning Areas -
- Hands-on Experience with Computer Vision
- Familiar With : NLP, GANs
- Computer Vision Areas -
- Hands-on Experience with Image Classification
- Working Knowledge Of Image Segmentation, Object Detection
- Version Control Tools -
- Git
- GitHub
- Web Framework -
- Familiar with Flask
- Tools/IDEs -
- Jupyter Notebook
- Google Colab
- PyCharm