代做COMP5318/COMP4318 – Machine Learning and Data Mining Semester 1, 2025 Assignment 1代写Python编程

COMP5318/COMP4318 – Machine Learning and Data Mining

Semester 1, 2025

Assignment 1: Rice Classification

Key information

Deadlines

Submission: 11:59pm, 7 April 2025 (Monday week 7)

Late submissions policy

Late submissions are allowed for up to 3 days late. A penalty of 5% per day late will apply. Assignments more than 3 days late will not be accepted (i.e. will get 0 marks). The day cut-off time is 11:59pm.

Marking

This assignment is worth 10 marks = 10% of your final mark.

Please see the marking rubric on Canvas. Most of the marks are for your code – it will be marked for correctness. A few marks will be allocated for code style. – meaningful variable names and comments, and also for reflection.

We will run your code. If it doesn’t run, you will get 0 marks for the parts that don’t run.

Submission

The assignment must be written in Python in the Jupyter Notebook environment. A Jupyter Notebook template is provided. Your implementation should use the same suite of libraries that we have used during the tutorials, such as scikit-learn, numpy and pandas. The submission is in Canvas.

Groups

The assignment should be completed in pairs (groups of 2 students). No more than 2 students are allowed.

Firstly (as announced on Ed in week 2), you need to create a group and register it in Canvas->People. There are 2 tabs (group sets) in People: A1 group-set1 and A1 group-set2. You must register your group under one of these group-sets, not both. We had to create 2 group-sets because Canvas has a limitation for the number of groups under one tab and our course is big.

Choose one of the empty groups listed under one of these tabs and join it. You and your partner must join the same group.

Submission instructions:

•    When you are ready to submit your assignment, you need to submit it on behalf of the group in the corresponding submission box, not individually. Only one student from the group needs to submit, not both. When this is done, both students will be able to see the submission, and later see the mark (when the assignment is marked).

•    You need to submit two versions of your code: ipynb and pdf.

•    Make sure that you submit  in the correct submission  box! There may be penalties for incorrect submission. In summary:

o If you  have  registered  your  group  under  "A1  group-set1",  submit  your  ipynb  code  in "Submission:  Assignment   1  ipynb  for A1 group-set1"   and  your   pdf   in  "Submission: Assignment 1 pdf for A1 group-set1"

o  Similarly,  if you  have  registered your group  under "A1 group-set2", submit your ipynb code in "Submission: Assignment 1 ipynb for A1 group-set2" and your pdf in "Submission: Assignment 1 pdf for A1 group-set2"

It is important to follow the submission instructions carefully as otherwise your mark may not be recorded correctly.

File names and student names

•    The file names should include your group number, e.g. a1-set1-group3.ipynb (.pdf), if you are from group 3 in group-set1 or a1-set2-group180.ipynb (.pdf) if you are from group 180 in group-set2.

•    In the Jupyter Notebook, include your group name and your SIDs (as shown in the template). Do not include your name as the marking is anonymous.

A pdf of your Jupyter notebook can be generated using File>Download as>PDF or Print Preview > Save as PDF or File->Download as >HTML, then open the html file and save it as pdf.

Task

In this assignment you will implementing multiple classification algorithms and apply them to a real dataset. First, you will pre-process the dataset by replacing missing values and normalising the values. Second, you will implement and evaluate the performance of multiple classification algorithms: K-Nearest Neighbour, Logistic  Regression,  Naïve  Bayes, Decision Tree, Support Vector  Machine,  Bagging, AdaBoost, Gradient Boosting and Random Forest, using the stratified 10-fold cross-validation method. Third, you will apply grid search to find the best parameters for some of these classifiers. Forth, you need to write a short reflection.

1. Data loading, pre-processing and printing

The dataset for this assignment is the Rice dataset which describes two varieties of rice grains. It contains 1400 examples, described with 7 features which have been extracted from rice grain images. There are two classes – class1, corresponding to  rice variety “cammeo” and class2 – to  rice variety “osmancik” . The original     dataset     is      bigger     and     available      from     the     UCI      Machine     Learning      repository: https://archive.ics.uci.edu/dataset/545/rice+cammeo+and+osmancik .

For this assignment, you should use the rice-final2.csv dataset provide on Canvas. It is a smaller and slightly modified version of the original dataset. This file includes the attribute (feature) headings and each row corresponds to one grain example. Missing attributes in the dataset are recorded with a ‘?’ .

You will need to pre-process the dataset, before you can apply the classification algorithms. Three types of pre-processing are required: filling in the missing values, normalisation and changing the class values. After this is done, you need to print the first 10 rows of the pre-processed dataset.

1.    Filling in the missing attribute values - The missing attribute values should be replaced with the mean value of the column using sklearn.impute.SimpleImputer.

2.    Normalising the data - Normalisation of each attribute should be performed using a min-max scaler to normalise the values between [0,1] with sklearn.preprocessing.MinMaxScaler.

3.    Changing  the  class  values  -  The  classes class1 and class2 should  be  changed to 0 and 1 respectively.

4.    Print the first 10 rows of the pre-processed dataset. The feature values should be formatted to 4 decimal places using .4f, the class value is an integer. A function print_data has been provided in the template to help you achieve this.

For example, you program should print the first 10 rows in this format (10 rows not 3, and the actual numbers, these are random numbers):

0.2861,0.2712,0.2086,0.4853,0.4296,0.2900,0.2247,1 0.5004,0.5781,0.5512,0.4854,0.7636,0.5374,0.7630,0 0.8077,0.8890,0.8694,0.6393,0.8541,0.8079,0.0747,0

Please note that we will test your code with another dataset, and your pre-processing should be written with this in mind. See the “Marking Criteria” section for more detail.

2. Defining functions for the classification algorithms

Part 1: Cross-validation without parameter tuning

You will now apply multiple classifiers to the processed dataset, in particular: Logistic Regression, Naïve Bayes, Decision Tree, Bagging, Ada  Boost and Gradient  Boosting. All classifiers should  use the sklearn modules from the tutorials. All random states in the classifiers should be set to random_state=0.

You need to evaluate the performance of these classifiers using 10-fold stratified cross-validation from sklearn.model_selection.StratifiedKFold with these options:

cvKFold=StratifiedKFold(n_splits=10, shuffle=True, random_state=0)

You will need to pass cvKFold (the stratified folds) as an argument when calculating the cross-validation accuracy, not cv=10 as in the tutorials. This ensures that random_state=0.

For each classifier, write a function that accepts the required input and returns the average cross-validation score:

def exampleClassifier(X, y, [options]):

…

return scores.mean()

where X contains the attribute values and y contains the class (as in the tutorial exercises).

Specifically, the headers of the functions for the classifiers are given below and provided in the template. You must not change them; your task is to write the body of these functions.

Logistic Regression

def logregClassifier(X, y)

…

return scores.mean()

It should use LogisticRegression from sklearn.linear_model.

Naïve Bayes

def nbClassifier(X, y)

…

return scores.mean()

It should use GaussianNB from sklearn.naive_bayes

Decision Tree

def dtClassifier(X, y)

…

return scores.mean()

It should use DecisionTreeClassifier from sklearn.tree, with information gain (the entropy criterion)

Ensembles: Bagging, Ada Boost and Gradient Boosting

def bagDTClassifier(X, y, n_estimators, max_samples, max_depth)

…

return scores.mean()

def adaDTClassifier(X, y, n_estimators, learning_rate, max_depth)

…

return scores.mean()

def gbClassifier(X, y, n_estimators, learning_rate)

…

return scores.mean()

These functions should implement Bagging, Ada Boost and Gradient Boosting using BaggingClassifier, AdaBoostClassifier and GradientBoostingClassifier from sklearn.ensemble, respectively. Bagging and Ada Boost should combine decision trees and use information gain.

Part 2: Cross-validation with parameter tuning

For two other classifiers, SVM and Random Forest, we would like to find the best parameters using grid search with 10-fold stratified cross-validation (GridSearchCV in sklearn).

The    data    should    be    split    into    training    and    test    subsets    using    train_test_split  from sklearn.model_selection with   stratification  and  random_state=0  (as  in  the  tutorials   but   with random_state=0).

You will need to pass cvKFold (the stratified folds) as an argument to GridSearchCV, not cv=10 as in the tutorials. This ensures that random_state=0.

Write the following functions:

K-Nearest Neighbour

def bestKNNClassifier(X, y)

…

return (appropriate values so that the required printing can be done)

It should use the KNeighborsClassifier from sklearn.neighbors.

The grid search should consider the following values for the parameters k(n_neighbors) and p:

k = {1, 3, 5, 7, 9}

p = {1, 2}

The function should return appropriate values, so that best parameters found, the best cross-validation accuracy and the test set accuracy can be printed when calling this function, see the next section.

SVM

def bestSVMClassifier(X,y)

…

return (appropriate values so that the required printing can be done)

It should use SVC from sklearn.svm with kerneI set to (rbf,.

The grid search should consider the following values for the parameters C and gamma:

C = {0.01, 0.1, 1, 5}

gamma = {0.01, 0.1, 1, 10}

The function should return appropriate values, so that best parameters found, the best cross-validation accuracy and the test set accuracy can be printed when calling this function, see the next section.

Random Forest

def bestRFClassifier(X,y)

It should use RandomForestClassifier from sklearn.ensemble with information gain and max_features set to (sqrt,.

The grid search should consider the following values for the parameters n_estimators and max_leaf_nodes: n_estimators = {10, 30, 60, 100}

max_leaf_nodes = {6, 12}

The function should return appropriate values, so that best parameters found, the best cross-validation accuracy, the test set accuracy, the macro average F1 score, and the weighted average F1 score can be printed when calling this function, see the next section.

3. Running the classifiers and printing the results

Run the classifiers from the previous section on the pre-processed dataset and print the results.

For Part 1 of this assignment, set the parameters as follows (this is already done for you in the template):

#Bagging bag_n_estimators = 50 bag_max_samples = 100 bag_max_depth = 5 #AdaBoost ada_n_estimators = 50 ada_learning_rate = 0.5 ada_bag_max_depth = 5 #GB gb_n_estimators = 50 gb_learning_rate = 0.5

The printing should look like this but with the correct numbers (these are random numbers):

LR average cross-validation accuracy: 0.8123 NB average cross-validation accuracy: 0.7543 DT average cross-validation accuracy: 0.6345 Bagging average cross-validation accuracy: 0.8765 AdaBoost average cross-validation accuracy: 0.7165 GB average cross-validation accuracy: 0.9054 KNN best k: 7 KNN best p: 2 KNN cross-validation accuracy: 0.7853 KNN test set accuracy: 0.5991 SVM best C: 0.0100 SVM best gamma: 10.0000 SVM cross-validation accuracy: 0.8676 SVM test set accuracy: 0.8098 RF best n_estimators: 10 RF best max_leaf_nodes: 16 RF cross-validation accuracy: 0.8600 RF test set accuracy: 0.8321 RF test set macro average F1: 0.8123 RF test set weighted average F1: 0.8261

Format all numbers to 4 decimal places using .4f, except k, p, n_estimators and max_leaf_nodes which should be formatted as integers.

4. Reflection

In the last cell of the template, each student needs to write a short reflection (1 paragraph):

What was the most important thing that you have learnt throughout this assignment?

Marking Criteria

The marking rubric is provided in Canvas.

Please note that we will test your program on another dataset. It will have the same format as the rice dataset but a different number of features and examples. Hence, do not hard-code the number of features and examples - do not set them to 1400 and 7 as in the rice dataset. You may assume the class value will be in the last column and there will be two classes (class1 and class2) as in the rice dataset.

To test your code before submission, we have made available another dataset (test-before.csv), with the correct results.