dnes, 6:48
I'm helping a few juniors who are stuck at the topic stage, and I realized there is no single right way to pick one. Some people start with a dataset and see what story it tells. Others begin with a problem from daily life and work backwards to a model.
From what I have seen, machine learning projects for final year go wrong in three common ways. First, the topic is too big, so nothing gets finished. Second, the data is hard to find or too messy to clean in time. Third, the student can run the code but cannot explain why the model works, which shows up quickly in the viva.
The projects that seem to go best are small and clear. Think disease prediction, spam or fake news detection, crop yield forecasting, or sentiment analysis. They have available datasets, clear metrics, and enough room to add something of your own, like better preprocessing or a comparison of two models.
A few questions for everyone here:
#MachineLearning #FinalYearProjects #CSEStudents #ECEProjectKart
From what I have seen, machine learning projects for final year go wrong in three common ways. First, the topic is too big, so nothing gets finished. Second, the data is hard to find or too messy to clean in time. Third, the student can run the code but cannot explain why the model works, which shows up quickly in the viva.
The projects that seem to go best are small and clear. Think disease prediction, spam or fake news detection, crop yield forecasting, or sentiment analysis. They have available datasets, clear metrics, and enough room to add something of your own, like better preprocessing or a comparison of two models.
A few questions for everyone here:
- What topic did you choose, and why?
- What was the hardest part: data, training, or the report?
- Did your evaluators care more about accuracy or about your understanding?
- What would you do differently if you started again?
#MachineLearning #FinalYearProjects #CSEStudents #ECEProjectKart