代写DATA 201 – Assignment 4代做留学生Python程序

DATA 201 - Assignment 4

Total marks: 20

Due date: 11:59 p.m., Thursday, September 26, 2024.

Submit code and outputs in a single Jupyter notebook file. Do not expect the marker to rerun your code in order to get the outputs.

Your task is to develop a machine learning model to predict the market value of houses in a European city, using data (file data.csv) on various characteristics of each property.

The data description is provided below.

id: a unique identifier for each house

zipcode: each house is associated with a zipcode that corresponds to its location

lot-len: the plot length in meters where the house is built

lot-width: the plot width in meters where the house is built

lot-area: the total plotarea on which the house is built

house-area: the house's living area measured in square meters

garden-size: the garden's size measured in square meters

balcony: the total number of balconies in the house

x-coor: the x-coordinate representing the house's location

y-coor: they-coordinate representing the house's location

buildyear: the year in which the house was built

bathrooms: the total number of bathrooms in the house

value: the market value of the house (in Euros)

Requirements:

●   Use root mean square error (RMSE) as the evaluation metric. [2 marks]

●   Load the dataset, determine the target column, remove irrelevant variables (if any), and use function train_test_split with random_state=1 to split the data into two sets: atraining set (80%) and a test set (20%). [3 marks]

●   Explore the training set to gain insights. [2 marks]

●   Select one machine learning model, train it, optimise it (e.g., add pre-processing

transformers, perform hyper-parameter tuning, etc.), and estimate the performance of the model. [9 marks]

●   Test the final model on the test set and report the RMSE and at least two other

evaluation metrics (e.g., mean absolute percentage error (MAPE), R2-score, etc.). [3 marks]

●   Include a discussion at the end of your notebook (about what you have learned, difficulties, what has worked and not worked, future directions, etc.). [1 mark]

Notes:

-     Write your name and student ID at the beginning of your notebook. After completing your work, use menu item Kernel => Restart & Run All in Jupyter, then submit your notebook file.

-     You can use any public Python package.

-     The requirements above have no order that you have to follow.

-     Use your own assumptions and judgement if you are unsure about any information in the dataset. However, remember to mention it in the discussion.

-     Try to write functions for all data transformations you apply, try feature engineering

(e.g., creating new features), and try to automate all the steps as much as possible (e.g., using pipeline and data transformers, etc.). You may have bonus marks for this; however, your total mark will not exceed 20.