Skip to main content
iRubric: Final Project: Animal Classification rubric

iRubric: Final Project: Animal Classification rubric

find rubric

edit   print   share   Copy to my rubrics   Bookmark   test run   assess...   delete   Do more...
Final Project: Animal Classification 
Summary Paper Rubric
Rubric Code: W223W2C
By Werner97
Ready to use
Public Rubric
Subject: Computers  
Type: Project  
Grade Levels: Graduate

Powered by iRubric Criterion
  Poor

0 pts

Fair

1 pts

Good

2 pts

Excellent

3 pts

Data Analysis
9 pts

0-3-6-9
Poor

- None
Fair

- Data was only loaded into training, validation and test sets.

- No samples are plotted.
Good

- Data was only loaded into training, validation and test sets and samples plotted.

- No commentary.
Excellent

- Data was only loaded into training, validation and test sets and samples plotted.

- With insightful commentary.
Data Augmentation
12 pts

0-4-8-12
Poor

- None
Fair

- Data was augmented with one technique (flipped, rotated, shifted, zoomed, photometric distortions, etc).

- Little to no commentary.
Good

- Data was augmented with one technique (flipped, rotated, shifted, zoomed, photometric distortions, etc).


- Identified that data augmentation is required with commentary supplied as to why. (Training set size is too small, helps combat overfitting, improves model robustness, etc)
Excellent

- Data was augmented using more than one technique (flipped, rotated, shifted, zoomed, photometric distortions, etc).


- Identified that data augmentation is required with commentary supplied as to why. (Training set size is too small, helps combat overfitting, improves model robustness, etc)
Data Preparation
18 pts

0-6-12-18
Poor

- None
Fair

- Data reshaped using centering, normalization or standardization.

- no commentary.
Good

- Data reshaped using centering, normalization or standardization.

- Only one of the data preprocessing techniques were acknowledged.

- Motivations as to why. (model input range limitations, memory issues, compensate for RGB values, etc)
Excellent

- Data reshaped using centering, normalization or standardization.

- More than one type of data preprocessing technique was considered.

- Motivation for the choice of technique.
Model Choice
15 pts

0-5-10-15
Poor

- Only one model considered.

- No motivation.
Fair

- Only one model considered.

- Motivation for choice of model.
Good

- Only one model chosen but multiple were considered.

- Motivation for choice of final model
Excellent

- More than one type of model (or different configurations of the same type i.e. different structured NNs) chosen.

- Motivations as to why.
Model Training
21 pts

0-7-14-21
Poor

- Model(s) only trained.

- No commentary on training results.

- No commentary for choice of loss metrics, no graphs, etc.
Fair

- Model(s) trained

- Commentary on training results. (loss criteria, overfitting, how well the model fairs, etc)

- No commentary for choice of loss metrics, learning rates, selection of optimizers, etc
Good

- Model(s) trained

- Commentary on training results. (loss criteria, overfitting, how well the model fairs, etc)

- Commentary for choice of loss metrics, learning rates, selection of optimizers, etc

- No graphs or visualization of training.
Excellent

- Model(s) trained

- Commentary on training results. (loss criteria, overfitting, how well the model fairs, etc)

- Commentary for choice of loss metrics, selection of optimizers, etc

- Training visualized through graphs.(confusion matrix, roc curve, accuracy estimates vs epoch, etc)
Model Tuning
18 pts

0-6-12-18
Poor

- No parameters are tuned for.

- No discussion regarding hyper-parameters.
Fair

- Hyper-parameters were tuned and evaluated.

- No discussion regarding hyper-parameters.
Good

- Hyper-parameters were tuned and evaluated on the validation set.

- Commentary on the different hyper-parameters.
Excellent

- Hyper-parameters were tuned and evaluated on the validation set.

- Commentary on the different hyper-parameters.

- Performance of different values of hyper-parameters visualized through graphs.

- Choice of tuning technique justified (grid search, etc).
Model assessment
21 pts

0-7-14-21
Poor

- No test results shown.
Fair

- Results shown through tabular data and/or graphs. (confusion matrix, ROC curve, etc)

- No discussion
Good

- Results portrayed through tabular data and/or graphs. (confusion matrix, ROC curve, etc)

- Commentary regarding the choice of performance metric.
Excellent

- Results portrayed through tabular data and/or graphs. (confusion matrix, ROC curve, etc)

- Commentary regarding performance of the model. (Different models compared)

- Commentary regarding the choice of performance metric. (Accuracy, cross entropy loss, etc)

- Discussion on how performance could be increased further.
Final Model Performance
18 pts

0-6-12-18
Poor

0-50 % Accuracy
Fair

50-70 % Accuracy
Good

70-90 % Accuracy
Excellent

90-100 % Accuracy
Presentation
15 pts

0-5-10-15
Poor

- No commentary.

- Graphs are incomplete (axes are not named, no titles).
Fair

- Little commentary throughout.

- Graphs are incomplete (axes are not named, no titles).
Good

- Commentary is sufficient.

- Graphs are incomplete (axes are not named, no titles).
Excellent

- Commentary is sufficient as well as insightful.

- Graphs are well presented (axes are labelled, graphs are titled).

- Notebook flows neatly.






Discuss this rubric




Do more with this rubric:
Preview

Preview

Preview this rubric.
Edit

Edit

Modify this rubric.
Copy

Copy

Make a copy of this rubric and begin editing the copy.

Print

Print

Show a printable version of this rubric.
Categorize

Categorize

Add this rubric to multiple categories.
Bookmark

Bookmark

Bookmark this rubric for future reference.
Assess
Test run

Test run

Test this rubric or perform an ad-hoc assessment.
Grade

Grade

Build a gradebook to assess students.
Collaborate

Collaborate

Apply this rubric to any object and invite others to assess.
Share
Publish

Publish

Link, embed, and showcase your rubrics on your website.
Email

Email

Email this rubric to a friend.
Discuss

Discuss

Discuss this rubric with other members.
 

Do more with rubrics than ever imagined possible.

Only with iRubrictm.


Copyright © 2024 Reazon Systems, Inc.  All rights reserved.
n16