BMS5010讲解 、辅导 Python/Java程序
2024/25 Semester B - BMS5010 - Assignment 3
Task:
1. Use Keras to implement a three-layer feedforward neural network with two hidden layers for binary
classification.
2. Select a dataset from sklearn.datasets and apply the implemented classifier to it.
3. Evaluate the model's performance on the selected dataset using all the metrics discussed in the lecture
slides.
4. Conduct experiments to analyze the impact of increasing the number of hidden layers from two to
three, using evaluation metrics to compare results.
5. Write a report detailing your code and evaluation results, including the following sections:
Introduction, Implementation, Results, Discussion, and References.
Assessment:
- You must submit both the Jupyter Notebook (5 marks) and the PDF report (10 marks) via
Canvas-Assignment by April 15th at 6:00 PM.
- The word count must be at least 1,500 words. References should be excluded from the word count.
- While there is no strict minimum number of references required, failure to properly cite sources
will result in a deduction of marks.
Notification:
- Your submitted report will undergo plagiarism detection and Artificial Intelligence-Generated
Content (AIGC) analysis using Turnitin service provided on CityU Canvas (Figure 1). The
results must meet the following thresholds:
1. Similarity Ratio < 10%
2. AI Ratio < 20%
- Failure to meet these requirements will be considered academic misconduct, resulting in a final
course grade of F (Unsatisfactory). Additionally, the case will be reported to BMS, SGS, and
ARRO.
- We will only consider the similarity ratio and AI ratio provided by the Turnitin service on CityU
Canvas (Figure 1) for assessment. Ratios generated from any third-party services will not be
accepted.
Figure 1