1 00:00:00,990 --> 00:00:05,860 So let's write the function to demonstrate how to find error. 2 00:00:05,910 --> 00:00:12,260 I'm gonna make a copy of the large project or rename this to find error 3 00:00:24,710 --> 00:00:25,730 and I'm gonna open it 4 00:00:43,620 --> 00:00:44,240 right. 5 00:00:44,250 --> 00:00:47,560 So this our last project we simply need to go to a. 6 00:00:47,610 --> 00:00:52,600 Simple neural network to see fall and implement the function. 7 00:00:52,650 --> 00:00:53,100 Right. 8 00:00:53,130 --> 00:01:01,590 And in the third indeed three school class we said we compute the error by an input multiply by weight 9 00:01:01,680 --> 00:01:03,590 minus expected value. 10 00:01:03,690 --> 00:01:10,500 So we're going to write a simple function that accept three argument it's gonna return the error. 11 00:01:10,520 --> 00:01:16,130 So I'll see Britain double over here and the name of the function. 12 00:01:16,350 --> 00:01:17,820 Find error. 13 00:01:18,230 --> 00:01:24,230 And I'm going to call this simple cause we're gonna have another one. 14 00:01:24,740 --> 00:01:29,300 Um the error we're dealing with currently we said it's simply expected value. 15 00:01:29,980 --> 00:01:31,230 Um. 16 00:01:31,770 --> 00:01:32,520 Right. 17 00:01:32,550 --> 00:01:35,750 So what I'm trying to say is we can write two functions. 18 00:01:35,910 --> 00:01:42,170 We can write a function that simply takes the Y had value which is the um the predicted value. 19 00:01:43,110 --> 00:01:49,260 And then another function that takes the input and then the weight separately in order to compute the 20 00:01:49,590 --> 00:01:50,670 Y heart from that. 21 00:01:51,030 --> 00:01:52,720 So let's write both functions. 22 00:01:52,830 --> 00:01:58,950 I'm gonna have one function on a simple and this one is gonna take just to argument the first argument 23 00:01:58,950 --> 00:02:01,670 is Why had we just the predicted value. 24 00:02:01,680 --> 00:02:10,740 And the second argument is why like this double why open close and we can simply return 25 00:02:13,290 --> 00:02:23,610 power F and what we want to return is why hearts minus Y and raced to the power to the reason we get 26 00:02:23,610 --> 00:02:34,780 in this is we've not included this C library so we've got to include Math thought h over here. 27 00:02:36,180 --> 00:02:36,990 Right. 28 00:02:37,050 --> 00:02:43,800 So this is our simple error function less find error or simple function let's write another function 29 00:02:44,340 --> 00:02:50,310 and this function would allow us to parse the input and weight separately so that we can compute the 30 00:02:50,310 --> 00:02:55,680 Y had the sources double find error 31 00:02:59,010 --> 00:03:03,340 and the first argument we can see double input. 32 00:03:03,930 --> 00:03:06,180 The second argument can be the weight 33 00:03:09,310 --> 00:03:15,170 on the last argument can be the expected value which is the wife argue 34 00:03:19,650 --> 00:03:25,970 we open and close and we simply want to return just like what we return here. 35 00:03:26,130 --> 00:03:31,620 But this time we have to compute y how to ourself ourselves I should say. 36 00:03:31,620 --> 00:03:40,650 So say input multiplied by weight in this gives us y hut and then we subtract expected value from this 37 00:03:42,150 --> 00:03:49,060 and then everything in brackets we raced to the post to then we get the error. 38 00:03:49,740 --> 00:03:53,540 So that's the same as this function here. 39 00:03:54,130 --> 00:03:54,490 Right 40 00:04:09,820 --> 00:04:10,660 let's see. 41 00:04:10,680 --> 00:04:15,510 It says over here the practice this practice for this one. 42 00:04:15,510 --> 00:04:17,940 This one is with this one. 43 00:04:17,940 --> 00:04:19,140 This one is with this one 44 00:04:25,060 --> 00:04:25,470 bright. 45 00:04:25,470 --> 00:04:29,330 The reason we get in this is we want to bring the function. 46 00:04:29,700 --> 00:04:30,090 Right. 47 00:04:30,180 --> 00:04:30,570 Okay. 48 00:04:30,600 --> 00:04:32,930 So let's expose these functions. 49 00:04:32,940 --> 00:04:35,430 I'm gonna copy this. 50 00:04:35,430 --> 00:04:36,280 Bring it to the dot. 51 00:04:36,310 --> 00:04:37,290 Each file over here 52 00:04:40,190 --> 00:04:43,050 put a semicolon here and the next one. 53 00:04:43,410 --> 00:04:47,890 Okay the next one has a different number of arguments. 54 00:04:48,030 --> 00:04:48,370 Easier. 55 00:04:48,390 --> 00:04:56,900 Copy it or copy this pasted over here like this right. 56 00:04:56,940 --> 00:04:59,640 So let's go to our main file and test it out. 57 00:05:01,710 --> 00:05:02,790 So let's see 58 00:05:05,710 --> 00:05:10,140 let's say we have some expected values here. 59 00:05:10,420 --> 00:05:13,350 I'm going to create a vector on a rate hold them. 60 00:05:13,600 --> 00:05:18,670 I'll say I'll just call them literally expected values. 61 00:05:18,670 --> 00:05:24,900 And the length of expected values is equal to the output land. 62 00:05:25,630 --> 00:05:28,350 And I can just see we expect. 63 00:05:28,390 --> 00:05:30,970 So this is basically the y values. 64 00:05:30,970 --> 00:05:35,730 These are the y values that we've sort of measured empirically. 65 00:05:36,130 --> 00:05:43,450 It's an empirically um derived value that we want to train the one year in work with. 66 00:05:43,660 --> 00:05:48,540 I know you understand this perhaps you've been talking about it is making it confusing. 67 00:05:48,730 --> 00:05:56,590 So I'm simply going to say I mean that is or meaning this is the same US y values just for those of 68 00:05:56,590 --> 00:05:58,390 you who could forget. 69 00:05:58,390 --> 00:06:04,210 So over here we already predict in the we already have sad prediction. 70 00:06:04,210 --> 00:06:13,180 So to compute side error we can simply do this so we can print outside error and we do it by find arrow 71 00:06:13,210 --> 00:06:20,920 we call our function and then the first argument is the um the predicted the predicted output side we 72 00:06:20,920 --> 00:06:26,760 pass this here and then the second argument is the expected output facade and the indices are the same. 73 00:06:27,190 --> 00:06:30,660 So sad index here is the same. 74 00:06:30,730 --> 00:06:36,050 You located in the same position in the expected values array. 75 00:06:36,430 --> 00:06:39,820 I'm gonna paste this here and change it for sake 76 00:06:44,010 --> 00:06:46,920 and over here we simply need to see sic 77 00:06:50,600 --> 00:07:00,590 sic prediction index and sick prediction index and we do it once more for this one here active 78 00:07:03,630 --> 00:07:09,240 copy this pasted over here then we can change the index 79 00:07:12,170 --> 00:07:19,850 patient over here and paste over here like this and then control us to safe or click here to build 80 00:07:26,080 --> 00:07:35,080 and click here to download onto the board once this is done I'm going to open territory and I'm going 81 00:07:35,080 --> 00:07:36,750 to select my com port. 82 00:07:36,850 --> 00:07:43,330 It's this one over here and I'm gonna press reset on my board and this is where we have we've got the 83 00:07:43,390 --> 00:07:50,100 arrows and we've got the predicted values as well so this all the risk and I shall see you later.