ABW505 Python & Predictive Analytics Final Exam: Ultimate Revision Guide

🎓 Official Exam Structure

Q	Points	Type	Coverage
Q1	20	Type 2	Python basics: objects, variables, IO, list/tuple, functions, loops, conditions
Q2	30	Type 3	Choose 3 of 5: decisions, repetition, boolean logic, list/tuple, functions
Q3	25	Type 1/3	Pandas, preprocessing, encoders, SVM, Random Forest
Q4	25	Type 1/3/4	Naive Bayes, decision tree, preprocessing (calculator allowed)

Type definitions:

Type 1: theory
Type 2: read code, compute output
Type 3: read output, write code
Type 4: calculations

🟢 Part 1: Python Core & Logic (Q1 & Q2)

Q1.1 Mutable Default Parameter Trap

def extend_list(val, list=[]):
    list.append(val)
    return list
 
list1 = extend_list(10)
list2 = extend_list(123, [])
list3 = extend_list('a')

Key idea: default mutable args are evaluated once at function definition. The default list is reused across calls.

Q1.2 Loop Control

x = 0
for i in range(5):
    if i == 2:
        continue
    if i == 4:
        break
    x += i
print(x)

Result: 4

Q1.3 Tuple Unpacking

data = (10, 20, 30, 40, 50)
a, *b, c = data

Key idea: *b collects the middle items into a list.

Q1.5 NumPy Matrix Tasks

Create a 4×4 matrix from 1–16
Create a length-10 zero vector, set the 7th value to 10
Create an 8×8 checkerboard

import numpy as np
 
matrix_4x4 = np.arange(1, 17).reshape(4, 4)
null_vector = np.zeros(10)
null_vector[6] = 10
checkerboard = np.zeros((8, 8), dtype=int)
checkerboard[1::2, ::2] = 1
checkerboard[::2, 1::2] = 1

🟠 Part 2: Data Science Foundations (Q3 & Q4)

Pandas Essentials

import pandas as pd
 
df = pd.read_csv("data.csv")
print(df.head())
print(df.tail())

Key ops: head(), tail(), info(), describe(), groupby().

Preprocessing & Encoding

ML models need numeric input
Use one-hot encoding for categorical variables
Split data into train/test before evaluation

KNN Quick Rules

Small dataset, low dimensional → KNN works well
Larger K → lower variance, higher bias
Always normalize/scale features

Naive Bayes & Decision Tree

Naive Bayes: fast, good for text; assumes independence
Decision Tree: interpretable, but can overfit
Random Forest: improves tree stability via bagging

✅ Final Tips

For Type 2 questions: show step-by-step execution
For Type 3 questions: return clean, minimal code
Always handle edge cases (empty list, invalid input)
Use midpoint method in any percentage change calculation