|
1 |
|
2 ============= |
|
3 Probability |
|
4 ============= |
|
5 .. contents:: |
|
6 :local: |
|
7 |
|
8 .. role:: def |
|
9 :class: def |
|
10 |
|
11 PMF |
|
12 === |
|
13 |
|
14 :def:`PMF` or :def:`probability mass function` or :def:`probability law` or |
|
15 :def:`probability discribuion` of discrete random variable is a function that |
|
16 for given number give probability of that value. |
|
17 |
|
18 To denote PMF used notations: |
|
19 |
|
20 .. math:: |
|
21 |
|
22 PMF(X = x) = P(X = x) = p_X(x) = P({ω ∈ Ω: X(ω) = x}) |
|
23 |
|
24 where :math:`X` is a random variable on space :math:`Ω` of outcomes which mapped |
|
25 to real number via :math:`X(ω)`. |
|
26 |
|
27 Expected value |
|
28 ============== |
|
29 |
|
30 :def:`Expected value` of PMF is: |
|
31 |
|
32 .. math:: |
|
33 |
|
34 E[X] = Σ_{ω∈Ω} Χ(x) * p(ω) = Σ_{x} x * p_X(x) |
|
35 |
|
36 We write :math:`a ≤ X ≤ b` for :math:`∀ ω∈Ω a ≤ X(ω) ≤ b`. |
|
37 |
|
38 If :math:`X ≥ 0` then :math:`E[X] ≥ 0`. |
|
39 |
|
40 if :math:`a ≤ X ≤ b` then :math:`a ≤ E[X] ≤ b`. |
|
41 |
|
42 If :math:`Y = g(X)` (:math:`∀ ω∈Ω Y(ω) = g(X(ω))`) then: |
|
43 |
|
44 .. math:: |
|
45 |
|
46 E[Y] = Σ_{x} g(x) * p_X(x) |
|
47 |
|
48 **Proof** TODO: |
|
49 |
|
50 .. math:: |
|
51 |
|
52 E[Y] = Σ_{y} y * p_Y(y) |
|
53 |
|
54 = Σ_{y∈ℝ} y * Σ_{ω∈Ω: Y(ω)=y} p(ω) |
|
55 |
|
56 = Σ_{y∈ℝ} y * Σ_{ω∈Ω: g(X(ω))=y} p(ω) |
|
57 |
|
58 = Σ_{y∈ℝ} y * Σ_{x∈ℝ: g(x)=y} Σ_{ω∈Ω: X(ω) = x} p(ω) |
|
59 |
|
60 = Σ_{y∈ℝ} y * Σ_{x∈ℝ: g(x)=y} p_X(x) |
|
61 |
|
62 = Σ_{y∈ℝ} Σ_{x∈ℝ: g(x)=y} y * p_X(x) |
|
63 |
|
64 = Σ_{x∈ℝ} Σ_{y∈ℝ: g(x)=y} y * p_X(x) |
|
65 |
|
66 = Σ_{x} g(x) * p_X(x) |
|
67 |
|
68 .. math:: |
|
69 |
|
70 E[a*X + b] = a*E[X] + b |
|
71 |
|
72 Variance |
|
73 ======== |
|
74 |
|
75 :def:`Variance` is a: |
|
76 |
|
77 .. math:: |
|
78 |
|
79 var[X] = E[(X - E[X])^2] = E[X^2] - E^2[X] |
|
80 |
|
81 :def:`Standard deviation` is a: |
|
82 |
|
83 .. math:: |
|
84 |
|
85 σ_Χ = sqrt(var[X]) |
|
86 |
|
87 Property: |
|
88 |
|
89 .. math:: |
|
90 |
|
91 var(a*X + b) = a²̇ · var[X] |
|
92 |
|
93 |
|
94 Total probability theorem |
|
95 ========================= |
|
96 |
|
97 Let :math:`A_i ∩ A_j = ∅` for :math:`i ≠ j` and :math:`∑_i A_i = Ω`: |
|
98 |
|
99 .. math:: |
|
100 |
|
101 p_X(x) = Σ_i P(A_i)·p_{X|A_i}(x) |
|
102 |
|
103 Conditional PMF |
|
104 =============== |
|
105 |
|
106 :def:`Conditional PMF` is: |
|
107 |
|
108 .. math:: |
|
109 |
|
110 p_{X|A}(x) = P(X=x | A) |
|
111 |
|
112 E[X|A] = ∑_x x·p_{X|A}(x) |
|
113 |
|
114 Total expectation theorem |
|
115 ========================= |
|
116 |
|
117 .. math:: |
|
118 |
|
119 E[X] = Σ_i P(A_i)·E[X|A_i] |
|
120 |
|
121 To prove theorem just multiply total probability theorem by :math:`x`. |
|
122 |
|
123 Joint PMF |
|
124 ========= |
|
125 |
|
126 :def:`Joint PMF` of random variables :math:`X_1,...,X_n` is: |
|
127 |
|
128 .. math:: |
|
129 |
|
130 p_{X_1,...,X_n}(x_1,...,x_n) = P(AND_{x_1,...,x_n}: X_i = x_i) |
|
131 |
|
132 Properties: |
|
133 |
|
134 .. math:: |
|
135 |
|
136 E[X+Y] = E[X] + E[Y] |
|
137 |
|
138 Conditional PMF |
|
139 =============== |
|
140 |
|
141 :def:`Conditional PMF` is: |
|
142 |
|
143 .. math:: |
|
144 |
|
145 p_{X|Y}(x|y) = P(X=x | Y=y) = P(X=x \& Y=y) / P(Y=y) |
|
146 |
|
147 So: |
|
148 |
|
149 .. math:: |
|
150 |
|
151 p_{X,Y}(x,y) = p_Y(y)·p_{X|Y}(x|y) = p_X(x)·p_{Y|X}(y|x) |
|
152 |
|
153 p_{X,Y,Z}(x,y,z) = p_Y(y)·p_{Z|Y}(z|y)·p_{X|Y,Z}(x|y,z) |
|
154 |
|
155 ∑_{x,y} p_{X,Y|Z}(x,y|z) = 1 |
|
156 |
|
157 Conditional expectation of joint PMF |
|
158 ==================================== |
|
159 |
|
160 :def:`Conditional expectation of joint PMF` is: |
|
161 |
|
162 .. math:: |
|
163 |
|
164 E[X|Y=y] = ∑_x x·p_{X|Y}(x|y) |
|
165 |
|
166 E[g(X)|Y=y] = ∑_x g(x)·p_{X|Y}(x|y) |
|
167 |
|
168 Total probability theorem for joint PMF |
|
169 ======================================= |
|
170 .. math:: |
|
171 |
|
172 p_X(x) = ∑_y p_Y(y)·p_{X|Y}(x|y) |
|
173 |
|
174 Total expectation theorem for joint PMF |
|
175 ======================================= |
|
176 .. math:: |
|
177 |
|
178 E[X] = ∑_y p_Y(y)·E[X|Y=y] |
|
179 |
|
180 Independence of r.v. |
|
181 ==================== |
|
182 |
|
183 r.v. :math:`X` and :math:`Y` is :def:`independent` if: |
|
184 |
|
185 .. math:: |
|
186 |
|
187 ∀_{x,y}: p_{X,Y}(x,y) = p_X(x)·p_Y(y) |
|
188 |
|
189 So if two r.v. are independent: |
|
190 |
|
191 .. math:: |
|
192 |
|
193 E[X·Y] = E[X]·E[Y] |
|
194 |
|
195 var(X+Y) = var(X) + var(Y) |
|
196 |
|
197 Well known discrete r.v. |
|
198 ======================== |
|
199 |
|
200 Bernoulli random variable |
|
201 ------------------------- |
|
202 |
|
203 :def:`Bernoulli random variable` with parameter :math:`p` is a random variable |
|
204 that have 2 outcomes denoted as :math:`0` and :math:`1` with probabilities: |
|
205 |
|
206 .. math:: |
|
207 |
|
208 p_X(0) = 1 - p |
|
209 |
|
210 p_X(1) = p |
|
211 |
|
212 This random variable models a trial of experiment that result in success or |
|
213 failure. |
|
214 |
|
215 :def:`Indicator` of r.v. event :math:`A` is function:: |
|
216 |
|
217 I_A = 1 iff A occurs, else 0 |
|
218 |
|
219 .. math:: |
|
220 |
|
221 P_{I_A} = p(I_A = 1) = p(A) |
|
222 |
|
223 I_A*I_B = I_{A∩B} |
|
224 |
|
225 .. math:: |
|
226 |
|
227 E[bernoulli(p)] = 0*(1-p) + 1*p = p |
|
228 |
|
229 var[bernoulli(p)] = E[bernoulli(p) - E[bernoulli(p)]] |
|
230 |
|
231 = (0-p)²·(1-p) + (1-p)²·p = p²·(1-p) + (1 - 2p + p²)·p |
|
232 |
|
233 = p² - p³ + p - 2·p² + p³ = p·(1-p) |
|
234 |
|
235 Discret uniform random variable |
|
236 ------------------------------- |
|
237 |
|
238 :def:`Discret uniform random variable` is a variable with parameters :math:`a` |
|
239 and :math:`b` in sample space :math:`{x: a ≤ x ≤ b & x ∈ ℕ}` with equal |
|
240 probability of each possible outcome: |
|
241 |
|
242 .. math:: |
|
243 |
|
244 p_{unif(a,b)}(x) = 1 / (b-a+1) |
|
245 |
|
246 .. math:: |
|
247 |
|
248 E[unif(a,b)] = Σ_{a ≤ x ≤ b} x * 1/(b-a+1) |
|
249 = 1/(b-a+1) * Σ_{a ≤ x ≤ b} x |
|
250 |
|
251 = 1/(b-a+1) * (Σ_{a ≤ x ≤ b} a + Σ_{0 ≤ x ≤ b-a} x) |
|
252 |
|
253 = 1/(b-a+1) * ((b-a+1)*a + (b-a)*(b-a+1)/2) |
|
254 |
|
255 = a + (b-a)/2 |
|
256 = (b+a)/2 |
|
257 |
|
258 |
|
259 .. math:: |
|
260 |
|
261 var[unif(a,b)] = E[unif²(a,b)] - E²[unif(a,b)] |
|
262 |
|
263 = ∑_{a≤x≤b} x²/(b-a+1) - (b+a)²/4 |
|
264 |
|
265 = 1/(b-a+1)·(∑_{0≤x≤b} x² - ∑_{0≤x≤a-1} x²) - (b+a)²/4 |
|
266 |
|
267 = 1/(b-a+1)·(b+3·b²+2·b³ - (a-1)+3·(a-1)²+2·(a-1)³)/6 - (b+a)²/4 |
|
268 |
|
269 = (2·b² + 2·a·b + b + 2·a² - a)/6 - (b+a)²/4 |
|
270 |
|
271 = (b - a)·(b - a + 2) / 12 |
|
272 |
|
273 .. NOTE:: |
|
274 |
|
275 From Maxima:: |
|
276 |
|
277 sum(i^2,i,0,n), simpsum=true; |
|
278 |
|
279 2 3 |
|
280 n + 3 n + 2 n |
|
281 --------------- |
|
282 6 |
|
283 |
|
284 factor(b+3*b^2+2*b^3 - (a-1)-3*(a-1)^2-2*(a-1)^3); |
|
285 |
|
286 2 2 |
|
287 (b - a + 1) (2 b + 2 a b + b + 2 a - a) |
|
288 |
|
289 factor((2*b^2 + 2*a*b + b + 2*a^2 - a)/6 - (b+a)^2/4), simp=true; |
|
290 |
|
291 (b - a) (2 - a + b) |
|
292 ------------------- |
|
293 12 |
|
294 |
|
295 Binomial random variable |
|
296 ------------------------ |
|
297 |
|
298 :math:`Binomial random variable` is a r.v. with parameters :math:`n` (positive |
|
299 integer) and p from interval :math:`(0,1)` and sample space of positive integers |
|
300 from inclusive region :math:`[0, n]`: |
|
301 |
|
302 .. math:: |
|
303 |
|
304 p_{binom(n,p)}(x) = n!/(x!*(n-x)!) p^x p^{n-x} |
|
305 |
|
306 Binomial random variable models a number of success of :math:`n` independent |
|
307 trails of Bernoulli experimants. |
|
308 |
|
309 .. math:: |
|
310 |
|
311 E[binom(n,p)] = E[∑_{1≤x≤n} bernoulli(p)] = ∑_{1≤x≤n} E[bernoulli(p)] = n·p |
|
312 |
|
313 var[binom(n,p)] = var[∑_{1≤x≤n} bernoulli(p)] = ∑_{1≤x≤n} var[bernoulli(p)] = n·p·(1-p) |
|
314 |
|
315 Geometric random variable |
|
316 ------------------------- |
|
317 |
|
318 :def:`Geometric random variable` is a r.v. with parameter :math:`p` from |
|
319 half open interval :math:`(0,1]`, sample space is all positive numbers: |
|
320 |
|
321 .. math:: |
|
322 |
|
323 p_{geom(p)}(x) = p (1-p)^(x-1) |
|
324 |
|
325 This random variable models number of tosses of biased coin until first success. |
|
326 |
|
327 .. math:: |
|
328 |
|
329 E[geom(p)] = ∑_{x=1..∞} x·p·(1-p)^(x-1) |
|
330 |
|
331 = p·∑_{x=1..∞} x·(1-p)^(x-1) |
|
332 |
|
333 = p/(1-p)·∑_{x=0..∞} x·(1-p)^x |
|
334 |
|
335 = p/(1-p)·(1-p)/(1-p - 1)² = p/p² = 1/p |
|
336 |
|
337 .. NOTE:: |
|
338 |
|
339 Maxima calculation:: |
|
340 |
|
341 load("simplify_sum"); |
|
342 simplify_sum(sum(k * x^k, k, 0, inf)); |
|
343 Is abs(x) - 1 positive, negative or zero? |
|
344 negative; |
|
345 Is x positive, negative or zero? |
|
346 positive; |
|
347 Is x - 1 positive, negative or zero? |
|
348 negative; |
|
349 x |
|
350 ------------ |
|
351 2 |
|
352 x - 2 x + 1 |