Task: Dataset Selection and Neural Network Experimentation
https://www.kaggle.com/datasets/iamsouravbanerjee/…
Task 6 Task: Dataset Selection and Neural Network Experimentation Deadline for Presentation: Thursday, April 25th Team Members: Maximum three allowed Objective: The objective of this task is to enable trainees to select a classification dataset from either Kaggle or Hugging Face, build and compare several neural network architectures, perform hyperparameter tuning, and report on overfitting. Trainees will present their findings in a concise presentation to their peers within a limited timeframe. 1. Dataset Selection (15 marks): • Choose a classification dataset from either Kaggle or Hugging Face. • Ensure the dataset is approved by the trainer and meets the following criteria: o Number of Rows: At least 10,000 rows of data are generally considered suitable o Number of Categorical Columns: A minimum of 2 categorical columns is desirable o Total Number of Columns: at least 7, including both numerical and categorical features. o Offers a clear classification task with labeled data. o Provides features that are relevant and appropriate for the classification task. 2. Neural Network Experimentation (40 marks): • Experiment with building and comparing several neural network architectures on the selected dataset. • Explore various aspects of neural network design, including different architectures, activation functions, regularization techniques, and optimization algorithms. • Perform hyperparameter tuning on the neural network models to optimize performance. • Report on the presence of overfitting and discuss strategies for mitigating it. 3. Presentation Preparation (15 marks): • Prepare a presentation consisting of a maximum of 10 slides. • Ensure the presentation covers the following aspects: o Introduction to the selected dataset, including its source, size, and classification task. o Overview of the neural network architectures experimented with, including their design and hyperparameters. o Comparison of the performance of different neural network architectures and their hyperparameter-tuned versions. o Discussion of overfitting in the neural network models and strategies employed to address it. o Summary of findings and conclusions drawn from the experimentation process. 4. Presentation Delivery (30 marks): • Present the findings to peers in the class within a time limit of 5 minutes. • Deliver the presentation clearly and effectively, addressing all key points outlined in the presentation preparation. • Engage with peers by encouraging questions and discussion regarding the presented findings. Total Marks: 100 Note: • Seek approval from the trainer before finalizing the dataset for experimentation. • Collaboration, knowledge sharing, and feedback from peers and the trainer are encouraged throughout the experimentation and presentation process. • The presentation aims to provide a concise summary of the dataset, experimentation process, and findings related to neural network experimentation.
Collepals.com Plagiarism Free Papers
Are you looking for custom essay writing service or even dissertation writing services? Just request for our write my paper service, and we'll match you with the best essay writer in your subject! With an exceptional team of professional academic experts in a wide range of subjects, we can guarantee you an unrivaled quality of custom-written papers.
Get ZERO PLAGIARISM, HUMAN WRITTEN ESSAYS
Why Hire Collepals.com writers to do your paper?
Quality- We are experienced and have access to ample research materials.
We write plagiarism Free Content
Confidential- We never share or sell your personal information to third parties.
Support-Chat with us today! We are always waiting to answer all your questions.