技术伦理与四种编程范式

data = [
    # (群体, 真实标签, 模型预测)
    ('A', 1, 1), ('A', 1, 1), ('A', 0, 0), ('A', 0, 0),
    ('A', 1, 1), ('A', 0, 0), ('A', 1, 1), ('A', 0, 0),
    ('B', 1, 0), ('B', 1, 0), ('B', 0, 0), ('B', 0, 0),
    ('B', 1, 0), ('B', 0, 0), ('B', 1, 0), ('B', 0, 0),
]

correct = sum(1 for g, t, p in data if t == p)
print("总体准确率:", correct / len(data))

for group in ['A', 'B']:
    rows = [(t, p) for g, t, p in data if g == group]
    tp = sum(1 for t, p in rows if t == 1 and p == 1)
    fn = sum(1 for t, p in rows if t == 1 and p == 0)
    fnr = fn / (tp + fn) if (tp + fn) else 0
    print(f"群体 {group}: 假阴性率 = {fnr:.2f}")

from sqlite3 import connect

xs = [1, 2, 3, 4, 5, 6]

# ① 函数式：把计算组合起来
functional = sum(x * x for x in xs if x % 2 == 0)

# ② 面向对象：把状态和方法打包
class SumOfSquares:
    def __init__(self):
        self.total = 0
    def feed(self, v):
        self.total += v * v
        return self

s = SumOfSquares()
for x in xs:
    if x % 2 == 0:
        s.feed(x)
oop = s.total

# ③ 解释器：把表达式写成数据，再求值
def calc_eval(exp):
    if isinstance(exp, int):
        return exp
    op = exp[0]
    args = [calc_eval(a) for a in exp[1:]]
    if op == '+':
        return sum(args)
    if op == '*':
        r = 1
        for a in args:
            r *= a
        return r
    raise ValueError(op)

ast = ['+', ['*', 2, 2], ['*', 4, 4], ['*', 6, 6]]
interpreter = calc_eval(ast)

# ④ 声明式：只说「要什么」
c = connect(':memory:').cursor()
c.execute('create table xs (v int)')
for x in xs:
    c.execute('insert into xs values (?)', (x,))
declarative = c.execute('select sum(v * v) from xs where v % 2 = 0').fetchone()[0]

print("函数式  :", functional)
print("面向对象:", oop)
print("解释器  :", interpreter)
print("声明式  :", declarative)