Naive Bayes Classifier | Part 5 | Problem based upon Bayes Theorem — Transcript

Full transcript

1. 0:01So, now let's do a sample problem using
2. 0:03Bayes' theorem.
3. 0:05So, let's say there is a factory.
4. 0:10Let's say there is a factory. The
5. 0:12factory, let's say, manufactures these
6. 0:13markers, okay? And there are three
7. 0:16machines in this factory.
8. 0:18M1
9. 0:21M2
10. 0:24and M3.
11. 0:26So, whatever is manufactured
12. 0:34If all the data is available He didn't
13. 0:36know the production all right. What is
14. 0:38the 20%?
15. 0:40M1 banana So, I got there are 100
16. 0:43markers producing or manufacturing
17. 0:45getting manufactured.
18. 0:4720 out of those markers is made by M1.
19. 0:5230%
20. 0:57So,
21. 1:03He didn't know the banana.
22. 1:05What is my 5% manufacturing defect? What
23. 1:08happened?
24. 1:09You didn't know the banana. What is the
25. 1:113% manufacturing defect? What happened?
26. 1:13What is the banana?
27. 1:15What is the 1% manufacturing defect?
28. 1:17What happened? This is the data. Okay.
29. 1:19So, I got it. Out of 120 banana What is
30. 1:22the 5%? What happened? Okay. These are
31. 1:24the 8 club over. Yeah, so I got it. What
32. 1:27happened over? Okay, similarly
33. 1:30What is the 3 club over? You didn't know
34. 1:32the banana. What is the 8 club over?
35. 1:34Okay.
36. 1:38What is the question is this?
37. 1:41This is the question.
38. 1:43Question is
39. 1:45He imagined
40. 1:49When you randomly take marker with her
41. 1:52And What you want to tell her you would
42. 1:54defect it?
43. 1:57It is defective.
44. 1:58What is defective in the club?
45. 2:00You have to tell me the probability
46. 2:08I guess
47. 2:10There are three different machines in a
48. 2:12factory.
49. 2:19Or other than that randomly example is a
50. 2:21tell me
51. 2:29Okay, and this is the problem that you
52. 2:31have to solve.
53. 2:32Now, if you think carefully
54. 2:35First of all, you should always write to
55. 2:37the bus guy guy here.
56. 2:39So, I already know this P of M1
57. 2:43is
58. 2:45total
59. 2:46divided by total
60. 2:47so that is
61. 2:501 by 5
62. 2:5220 by 100 probability
63. 2:54Similarly,
64. 2:56probability of M2
65. 2:59is
66. 3:003 by 10
67. 3:02And similarly, probability of M3 is
68. 3:051 by 2
69. 3:10Okay.
70. 3:16P
71. 3:18A cheese defective let's say I'm
72. 3:21D signal defective only.
73. 3:24So, you defective only probability given
74. 3:29M1 You defective only probability given
75. 3:32item M2 you defective only probability
76. 3:34given item M3 so that
77. 3:36So, I guess we can write that
78. 3:37conditional probability so you can write
79. 3:40it like this. Probability of defective
80. 3:43given M1 is equal to 5 by 100 so you can
81. 3:48write it like 1 by 20.
82. 3:50Probability of D given M2
83. 3:55is
84. 3:56uh
85. 3:593 by 100.
86. 4:03And probability of
87. 4:05uh
88. 4:06M3 as in probability of D given
89. 4:10M3 is equal to 1 by 100.
90. 4:14Okay. So, you get all the probabilities.
91. 4:16So, you get all the questions we have.
92. 4:18Up to that important thing which is not
93. 4:21important. It is about that.
94. 4:22I mean randomly example draw here on the
95. 4:24defective example.
96. 4:25We need to find out the M3 from which it
97. 4:28was made and it is a probability here.
98. 4:30So, if you ask me, I would say I have to
99. 4:33find out the probability of
100. 4:35that item that we randomly draw drew out
101. 4:38of the sample is from M3
102. 4:42given it is defective.
103. 4:46Given it was defective, was M3 made here
104. 4:48or not, is the probability we need to
105. 4:50find out.
106. 4:50And by base theorem, you know this
107. 4:53that this will be equal to probability
108. 4:55of
109. 4:57uh sorry, probability of defective given
110. 5:02M3 multiplied by probability of
111. 5:06M3 divided by probability of
112. 5:09D.
113. 5:10I guess base theorem you have
114. 5:12probability of A given B is equal to
115. 5:15probability of B given A multiplied by P
116. 5:19of A divided by P of B.
117. 5:23This is base theorem, right?
118. 5:25I'm going to
119. 5:26get these three things to get you to
120. 5:28calculate the probability. And guess
121. 5:29what, some of these two things are
122. 5:30already here. You can totally see keep
123. 5:34you want to seize just take that.
124. 5:37Or you want to seize
125. 5:40you want to seize that.
126. 5:41There is only one thing that I need to
127. 5:43calculate and that is this part.
128. 5:48Probability of D.
129. 5:50So basically, I have to find out
130. 5:52uh
131. 5:53I have to find out the overall data make
132. 5:56my example draw
133. 5:58to see the effective probability here.
134. 6:00You want to show me I have to find out P
135. 6:02of B.
136. 6:04I'll put social media.
137. 6:05How can I find out P of defective? We
138. 6:07need to put the data in the total
139. 6:08percent mild effective and you can see
140. 6:10the data.
141. 6:12Social media.
142. 6:14Total defective mild out of the entire
143. 6:17batch
144. 6:18is basically coming from one of these
145. 6:20machines. So don't you think you can
146. 6:22write it like this?
147. 6:23Probability of
148. 6:25a defective piece is actually equal to
149. 6:29probability of
150. 6:31it being defective and coming from M1.
151. 6:37Right?
152. 6:38It's a lot of people I don't defective
153. 6:40here
154. 6:41or M1 set.
155. 6:43Right?
156. 6:44Plus probability of defective and coming
157. 6:47from M2 plus probability of defective
158. 6:51and coming from
159. 6:53M3.
160. 6:54Social media. It's so good English to
161. 6:55read in the media you will actually
162. 6:57understand the camera.
163. 6:59Total defective only got probability is
164. 7:01actually the summation of these three
165. 7:03individual probabilities.
166. 7:04Probability of a marker being defective
167. 7:07and coming from M1. That is one way of
168. 7:10getting a defective marker. Probability
169. 7:12of being defective and coming from M2.
170. 7:14That is the second way of getting a
171. 7:15defective marker and probability of
172. 7:17defective and coming from M3. This is
173. 7:19the third way of getting a defective
174. 7:20marker. So there are only three ways
175. 7:22that I can get a defective marker in the
176. 7:23something or whatever you have to leave
177. 7:25here.
178. 7:26It is I guess something you have to
179. 7:27learn.
180. 7:28Now, if you remember conditional
181. 7:31probability
182. 7:32uh we have to put out probability of A
183. 7:36given B is actually equal to P A
184. 7:39intersection B divided by P of B.
185. 7:43What if I do this? B.
186. 7:45What if I do this?
187. 7:51probability of
188. 7:54defective given
189. 7:57M1
190. 7:59multiplied by probability of
191. 8:02M1
192. 8:05plus probability of
193. 8:08B
194. 8:09given M2 multiplied by P of M2
195. 8:14plus probability of B given
196. 8:19M3 multiplied by probability of
197. 8:22M3 So I'm going to PD calculate with the
198. 8:25expression song and guess what?
199. 8:31or you know PD
200. 8:33will die or is the expression I just
201. 8:35want to solve and guess
202. 8:39using base theorem and conditional
203. 8:40probability you're able to solve a
204. 8:43difficult seeming a seemingly
205. 8:47so easy
206. 8:48Okay, so it will
207. 8:52I guess for sure solve the problem. So
208. 8:55thanks for watching.