Ôn Tập Tổng Hợp

1# === Variables & Types ===
2name = "MinAI"                  # str
3age = 25                        # int
4pi = 3.14                       # float
5is_active = True                # bool
6
7# === Collections ===
8fruits = ["apple", "banana"]    # list (mutable, ordered)
9coords = (10.5, 20.3)          # tuple (immutable)
10unique = {1, 2, 3}             # set (unique values)
11person = {"name": "A", "age": 25}  # dict (key-value)
12
13# === List Operations ===
14fruits.append("cherry")
15fruits.extend(["date", "fig"])
16fruits.pop()
17sorted_list = sorted(fruits)
18squares = [x**2 for x in range(10)]           # List comprehension
19evens = [x for x in range(20) if x % 2 == 0]  # Filtered
20
21# === Control Flow ===
22if score >= 90:
23    grade = "A"
24elif score >= 80:
25    grade = "B"
26else:
27    grade = "C"
28
29for item in fruits:
30    print(item)
31
32for i, item in enumerate(fruits):
33    print(f"{i}: {item}")
34
35# === Functions ===
36def calculate(a, b, operation="add"):
37    if operation == "add":
38        return a + b
39    return a - b
40
41# Lambda
42square = lambda x: x ** 2
43sorted_data = sorted(students, key=lambda s: s['gpa'], reverse=True)
44
45# *args, **kwargs
46def flexible(*args, **kwargs):
47    print(args)    # tuple
48    print(kwargs)  # dict
49
50# === String Methods ===
51text = "  Hello World  "
52text.strip()          # Remove whitespace
53text.lower()          # lowercase
54text.split()          # ['Hello', 'World']
55"-".join(["a","b"])   # "a-b"
56f"Name: {name}"       # f-string

1import numpy as np
2
3# === Array Creation ===
4a = np.array([1, 2, 3])
5zeros = np.zeros((3, 4))
6ones = np.ones((2, 3))
7rng = np.arange(0, 10, 2)         # [0, 2, 4, 6, 8]
8lin = np.linspace(0, 1, 5)        # 5 points from 0 to 1
9rand = np.random.randn(3, 3)      # Normal distribution
10
11# === Shape & Reshape ===
12a.shape                            # (3,)
13a.reshape(3, 1)                    # Column vector
14a.flatten()                        # 1D
15
16# === Indexing ===
17arr = np.array([[1,2,3],[4,5,6],[7,8,9]])
18arr[0, :]         # Row 0: [1, 2, 3]
19arr[:, 1]         # Col 1: [2, 5, 8]
20arr[arr > 5]      # Boolean: [6, 7, 8, 9]
21
22# === Math ===
23np.mean(a), np.std(a), np.median(a)
24np.sum(arr, axis=0)     # Column sums
25np.sum(arr, axis=1)     # Row sums
26np.dot(a, b)            # Dot product
27a @ b                   # Matrix multiply
28
29# === Broadcasting ===
30arr + 10                # Add 10 to all elements
31arr * np.array([1,2,3]) # Multiply each col

1import pandas as pd
2
3# === Create & Read ===
4df = pd.read_csv("file.csv")
5df = pd.read_excel("file.xlsx")
6df = pd.DataFrame({"a": [1,2], "b": [3,4]})
7
8# === Explore ===
9df.shape               # (rows, cols)
10df.info()              # Types + non-null counts
11df.describe()          # Statistics
12df.head()              # First 5 rows
13df.dtypes              # Column types
14df.isnull().sum()      # Missing counts
15
16# === Selection ===
17df['col']                       # Series
18df[['col1', 'col2']]            # Multiple cols
19df.loc[0, 'col']                # By label
20df.iloc[0, 0]                   # By position
21
22# === Filtering ===
23df[df['age'] > 30]
24df.query('age > 30 and city == "Hanoi"')
25df[df['city'].isin(['HN', 'HCM'])]
26
27# === Transform ===
28df['new'] = df['a'] + df['b']
29df['category'] = df['score'].apply(lambda x: 'A' if x > 90 else 'B')
30
31# === GroupBy ===
32df.groupby('city')['salary'].mean()
33df.groupby(['city', 'dept']).agg(
34    avg_salary=('salary', 'mean'),
35    count=('id', 'count')
36)
37
38# === Merge ===
39pd.merge(df1, df2, on='key', how='left')    # left/right/inner/outer
40pd.concat([df1, df2], axis=0)                # Stack vertically
41df.pivot_table(values='sales', index='region', columns='product', aggfunc='sum')
42
43# === Chain Operations (Best Practice) ===
44result = (
45    df
46    .dropna(subset=['col'])
47    .drop_duplicates()
48    .assign(new_col=lambda x: x['a'] + x['b'])
49    .query('value > 0')
50    .groupby('category')['value'].mean()
51)

1import matplotlib.pyplot as plt
2import seaborn as sns
3import plotly.express as px
4
5# === Seaborn (Statistical) ===
6sns.histplot(df['col'], kde=True)
7sns.boxplot(data=df, x='cat', y='num')
8sns.scatterplot(data=df, x='x', y='y', hue='group')
9sns.heatmap(df.corr(), annot=True, cmap='coolwarm')
10sns.pairplot(df, hue='target')
11
12# === Plotly (Interactive) ===
13px.scatter(df, x='x', y='y', color='cat', size='val',
14           hover_name='name', title='Scatter Plot')
15px.line(df, x='date', y='value', color='category')
16px.bar(df, x='cat', y='val', color='sub_cat', barmode='group')
17px.pie(df, values='val', names='cat', hole=0.4)
18px.histogram(df, x='col', nbins=30)
19
20# === Matplotlib (Foundation) ===
21fig, axes = plt.subplots(1, 2, figsize=(12, 5))
22axes[0].plot(x, y)
23axes[0].set_title('Line Plot')
24axes[1].bar(categories, values)
25axes[1].set_title('Bar Plot')
26plt.tight_layout()
27plt.show()

1# === Missing Values ===
2df.isnull().sum()
3df.fillna(df['col'].median(), inplace=True)    # Numeric
4df['cat'].fillna(df['cat'].mode()[0], inplace=True)  # Categorical
5df.dropna(subset=['important_col'])
6
7# === Duplicates ===
8df.duplicated().sum()
9df.drop_duplicates(subset=['col1', 'col2'], keep='first')
10
11# === Outliers (IQR) ===
12Q1, Q3 = df['col'].quantile([0.25, 0.75])
13IQR = Q3 - Q1
14mask = (df['col'] >= Q1 - 1.5*IQR) & (df['col'] <= Q3 + 1.5*IQR)
15df_clean = df[mask]
16
17# === Data Types ===
18df['date'] = pd.to_datetime(df['date'])
19df['price'] = df['price'].str.replace(',', '').astype(float)

1from sklearn.preprocessing import StandardScaler, OneHotEncoder
2from sklearn.pipeline import Pipeline
3from sklearn.compose import ColumnTransformer
4
5# === Encoding ===
6pd.get_dummies(df, columns=['cat_col'], drop_first=True)  # One-Hot
7df['edu_num'] = df['education'].map({'HS':0, 'BS':1, 'MS':2, 'PhD':3})  # Ordinal
8
9# === Scaling ===
10scaler = StandardScaler()
11X_train_scaled = scaler.fit_transform(X_train)
12X_test_scaled = scaler.transform(X_test)  # KHÔNG fit lại!
13
14# === Feature Creation ===
15df['revenue'] = df['price'] * df['quantity']
16df['month'] = df['date'].dt.month
17df['is_weekend'] = df['date'].dt.dayofweek.isin([5, 6]).astype(int)
18
19# === Feature Selection ===
20from sklearn.ensemble import RandomForestClassifier
21model = RandomForestClassifier().fit(X, y)
22importance = model.feature_importances_
23
24# === Pipeline (tránh Data Leakage!) ===
25preprocessor = ColumnTransformer([
26    ('num', Pipeline([('imputer', SimpleImputer(strategy='median')),
27                      ('scaler', StandardScaler())]), numeric_cols),
28    ('cat', Pipeline([('imputer', SimpleImputer(strategy='most_frequent')),
29                      ('ohe', OneHotEncoder(handle_unknown='ignore'))]), cat_cols)
30])
31full_pipe = Pipeline([('prep', preprocessor), ('model', RandomForestClassifier())])
32full_pipe.fit(X_train, y_train)