نبذة عني
Data Science and Artificial Intelligence professional with a strong foundation in Python, SQL, and data analysis. Skilled in writing SQL queries, data cleaning, visualization, and exploratory data analysis using tools su…
Data Science and Artificial Intelligence professional with a strong foundation in Python, SQL, and data analysis. Skilled in writing SQL queries, data cleaning, visualization, and exploratory data analysis using tools such as Pandas, NumPy, and Matplotlib. Familiar with machine learning algorithms and statistical concepts, with hands-on experience gained through academic and self-driven projects. Passionate about transforming data into actionable insights and eager to contribute to data-driven decision-making in a growth-oriented organization.
الخبرة
Python Developer | Data Science | Artificial Intelligence
المشاريع
CIFAR-10 Classification using CNN
This project demonstrates an image classification task using the CIFAR-10 dataset. I have compared a standard Artificial Neural Network (ANN) with a Convolutional Neural Network (CNN) to showcase the superiority of CNNs in handling spatial data (images).🚀 OverviewThe goal is to classify 32 \times 32 pixel color images into 10 distinct classes:Airplane, Automobile, Bird, Cat, Deer, Dog, Frog, Horse, Ship, Truck.🏗️ Model Architectures1. Artificial Neural Network (ANN)A simple multi-layer perceptron that flattens the input image.Layers: Input (Flatten) -> Dense (3000, ReLU) -> Dense (1000, ReLU) -> Output (10, Sigmoid).Observation: High memory usage and lower accuracy compared to CNN.2. Convolutional Neural Network (CNN)A specialized deep learning model for image processing.Layers: Conv2D -> MaxPooling -> Conv2D -> MaxPooling -> Flatten -> Dense -> Output.🛠️ Tech StackLanguage: PythonLibraries: TensorFlow, Keras, Matplotlib, NumPyDataset: CIFAR-10 (via Keras datasets)📊 Key ResultsNormalization: Pixel values were scaled from [0, 255] to [0, 1] to improve model convergence.Performance: The CNN model significantly outperformed the ANN by extracting spatial features through convolution filters.📋 How to RunClone the repository.Install dependencies: pip install tensorflow matplotlib numpy.Run the notebook or python script to train and evaluate the models.