Name: Anas Zafar
Type: User
Company: National University OF Computer And Emerging Sciences
Bio: Data Scientist with a Strong Foundation in Machine learning, Deep learning, Data Analysis, Building predictive models, Flask/Django Front-end Dev.
Twitter: anaszafar1108
Location: G11/3 Islamabad
Blog: linkedin.com/in/anas-zafar-02aa75194/
Anas Zafar's Projects
This project performs object recognition using CIFAR-10 and CNN/Random Forest models in Python. It preprocesses data, trains models, evaluates performance and compares results with the goal of high accuracy and feature importance understanding. The final output includes a comparison of the models and insights.
PSO feature selection improves classifier performance. Implemented in Jupyter Notebook with pandas, numpy, scikit-learn. PSO done from scratch. Results compared using accuracy, precision, recall, F1 score. Improves results compared to using all features. Can be applied to various classification problems.
Los Angeles City employee payroll data analyzed using hypothesis testing to answer 5 questions on bonus pay, department, overtime and health cost comparison. Results based on examination of provided dataset with empty fields handled.
This project aims to generate poetry in Roman Urdu using a dataset of poems by famous Urdu poets such as Allam Iqbal and Ghalib.
A dataset containing 12 features for predicting mortality by heart failure caused by Cardiovascular Diseases (CVDs) is analyzed. The dataset includes demographic information such as gender, age, and presence of risk factors like diabetes, anemia, high blood pressure, and smoking habits.
A C++ command line puzzle game where player rearranges pieces to complete a pattern. Input: puzzle size, piece state. Output: puzzle state after moves. ASCII characters represent pieces, no progress saved.
A research project on road condition detection using accelerometer and gyro sensors. The goal is to determine the accuracy of detecting road conditions by augmenting accelerometer readings with gyro sensor data.
This project is focused on scraping millions of emails dynamically from thousands of web pages automatically from the website [fredmiranda.com](https://fredmiranda.com/). The goal of this project is to create a dataset of email addresses that can be used for various purposes.
This project is aimed at scraping the data of university professors from Google Scholar, including their citation count, h-index, and other relevant information. The scraped data will be stored in a CSV file for further analysis.
The purpose of this project is to scrape faculty data (such as name, qualification, etc.) of the top 50 psychology universities and save it in a CSV file. This project uses Beautiful Soup and the requests library in Python to accomplish this task.
Develop machine learning model to detect diabetic retinopathy in eye images using k-mean clustering, image segmentation, and image classification with ANN, CNN, ML. Requires Numpy, OpenCV, Tensorflow, Sklearn.
This project develops an effective spell correction system for Roman Urdu using the Noisy Channel model. 4 components: language model, error model, candidate generation, and selection model. Suggests the most likely correction for a given incorrect word using probabilistic approach.
This project aims to develop a program that can accurately count the number of vehicles moving on a road in a given video and classify the traffic as high, low, or medium based on the count.
This project aims to develop a program that can perform transliteration between Roman and Urdu scripts using the `urduhack` library. Transliteration is the process of converting text from one script to another while preserving the pronunciation of the words.
A project locates the zebra crossing in an image and draws a bounding rectangle around it while counting the number of white strips.