1 00:00:00,133 --> 00:00:00,900 Hello my friends. 2 00:00:00,900 --> 00:00:03,900 Welcome back to the. Practical activities, this time. 3 00:00:03,900 --> 00:00:05,900 On the A club model still in. 4 00:00:05,900 --> 00:00:08,033 Association rule learning. 5 00:00:08,033 --> 00:00:08,333 All right. 6 00:00:08,333 --> 00:00:11,433 So this tutorial is actually going to be a quick one because. 7 00:00:11,633 --> 00:00:14,266 As you understood in Kairos Intuition. Lecture. 8 00:00:14,266 --> 00:00:17,700 Well the model is actually a simplified version 9 00:00:17,700 --> 00:00:21,466 of the Priory model because we only deal with the supports. 10 00:00:21,466 --> 00:00:22,600 And we don't even have. Rules. 11 00:00:22,600 --> 00:00:27,133 We only consider a set of products of which we analyze the support. 12 00:00:27,333 --> 00:00:30,166 So it's way simpler than the Priory model. 13 00:00:30,166 --> 00:00:33,700 And therefore that leads to the second reason why this will be quick. 14 00:00:33,966 --> 00:00:35,400 Really, if you have to choose. 15 00:00:35,400 --> 00:00:37,166 An association rule learning model to. 16 00:00:37,166 --> 00:00:39,200 Do association rule. Mining. 17 00:00:39,200 --> 00:00:42,500 I by no doubt. Recommend the Priory model. 18 00:00:42,733 --> 00:00:44,266 However, it's possible. 19 00:00:44,266 --> 00:00:45,000 That in some. 20 00:00:45,000 --> 00:00:48,133 Business problems you will only consider the support. 21 00:00:48,133 --> 00:00:51,066 You know you're only. Interested in doing. A support analysis. 22 00:00:51,066 --> 00:00:53,066 And therefore you might. Use it. 23 00:00:53,066 --> 00:00:54,433 In exceptional. Situations. 24 00:00:54,433 --> 00:00:57,200 But still, even with a priory you can do this. 25 00:00:57,200 --> 00:00:58,800 So that's why it's going to be quick. 26 00:00:58,800 --> 00:01:01,533 And besides, the way we're going to build. 27 00:01:01,533 --> 00:01:02,733 You know, in Python. 28 00:01:02,733 --> 00:01:04,700 Is by just adapting our A. 29 00:01:04,700 --> 00:01:05,233 Priory. 30 00:01:05,233 --> 00:01:08,433 Model so that we only consider the support right. 31 00:01:08,433 --> 00:01:10,400 Because indeed the class doesn't. 32 00:01:10,400 --> 00:01:11,833 Include any confidence or. 33 00:01:11,833 --> 00:01:13,800 Lift. Analysis. Okay. 34 00:01:13,800 --> 00:01:15,066 So let's do this quickly. 35 00:01:15,066 --> 00:01:17,833 This will. You know, give you an extra. Association. Rule. 36 00:01:17,833 --> 00:01:19,533 Learning model in the toolkit. 37 00:01:19,533 --> 00:01:20,966 So that's still good anyways. 38 00:01:20,966 --> 00:01:24,300 But really the focus should be on the Priory model. 39 00:01:24,900 --> 00:01:25,366 All right. 40 00:01:25,366 --> 00:01:27,533 So before we start let's make sure. 41 00:01:27,533 --> 00:01:29,000 Everyone here is on the same page. 42 00:01:29,000 --> 00:01:31,466 I give you the link to this folder right before this tutorial. 43 00:01:31,466 --> 00:01:32,966 So make sure to connect to it. 44 00:01:32,966 --> 00:01:33,566 And now. 45 00:01:33,566 --> 00:01:36,033 Let's all go into part five Association. 46 00:01:36,033 --> 00:01:36,966 Rule learning. 47 00:01:36,966 --> 00:01:42,466 Then section 29 Eeglab and Python where you will find two. 48 00:01:42,466 --> 00:01:43,533 Files the. 49 00:01:43,533 --> 00:01:47,000 Clamato in ipynb format and the same data. 50 00:01:47,000 --> 00:01:49,500 Set market basket. Optimization. 51 00:01:49,500 --> 00:01:51,300 Let's quickly remind the scenario. 52 00:01:51,300 --> 00:01:53,866 There is a shop. Owner in the south of France. 53 00:01:53,866 --> 00:01:57,066 Who would like to boost the sales of the shop, and therefore. 54 00:01:57,066 --> 00:01:59,400 He's trying to find the best association of. 55 00:01:59,400 --> 00:02:01,466 Products to sell to the. 56 00:02:01,466 --> 00:02:02,966 Customers in a deal. 57 00:02:02,966 --> 00:02:03,600 And the deal. 58 00:02:03,600 --> 00:02:06,466 That this owner. Has in mind is the following. 59 00:02:06,466 --> 00:02:09,233 Buy one product and get another one for free. 60 00:02:09,233 --> 00:02:10,800 So this time we're going to use. 61 00:02:10,800 --> 00:02:14,733 A Clat analysis to analyze the highest supports. 62 00:02:14,733 --> 00:02:16,466 Of combination of. Products. Here. 63 00:02:16,466 --> 00:02:18,300 Two products because the deal is buy. 64 00:02:18,300 --> 00:02:19,866 One, get another. One for free. 65 00:02:19,866 --> 00:02:21,133 So well, that's exactly. 66 00:02:21,133 --> 00:02:23,166 The same scenario as an a. Priori. 67 00:02:23,166 --> 00:02:25,166 And therefore let's directly. 68 00:02:25,166 --> 00:02:27,533 Get into the implementation. 69 00:02:27,533 --> 00:02:29,533 All right. Let's open. It. We're going to open it. 70 00:02:29,533 --> 00:02:32,766 With either Google Colaboratory or Jupyter Notebook. 71 00:02:32,766 --> 00:02:34,600 Choose your favorite. And 72 00:02:35,700 --> 00:02:38,200 now the notebook is opening. 73 00:02:38,200 --> 00:02:40,566 Soon it will be laying out. 74 00:02:40,566 --> 00:02:42,533 And in a second we should have it. 75 00:02:42,533 --> 00:02:43,433 There we go. 76 00:02:43,433 --> 00:02:44,700 All right. So loading it. 77 00:02:44,700 --> 00:02:47,366 Then laying it out and there. We go. We have it. 78 00:02:47,366 --> 00:02:49,466 So that's the. Implementation. 79 00:02:49,466 --> 00:02:50,500 As you will notice it. 80 00:02:50,500 --> 00:02:52,433 Is very similar to the a. 81 00:02:52,433 --> 00:02:53,866 Priori implementation because. 82 00:02:53,866 --> 00:02:55,533 Indeed the way I built. 83 00:02:55,533 --> 00:02:56,533 The model. 84 00:02:56,533 --> 00:03:00,566 Is just by adapting this binary package to the. 85 00:03:00,900 --> 00:03:03,633 Model by only considering the. Support. 86 00:03:03,633 --> 00:03:05,900 And I'm going to show you. Exactly how I. Did it. 87 00:03:05,900 --> 00:03:08,033 You know, I'm going to show you from. Scratch how I. 88 00:03:08,033 --> 00:03:09,100 Turned that. 89 00:03:09,100 --> 00:03:09,666 A priori. 90 00:03:09,666 --> 00:03:11,033 Implementation into. 91 00:03:11,033 --> 00:03:12,966 This new. Implementation. 92 00:03:12,966 --> 00:03:15,300 So here instead of, you know, creating. 93 00:03:15,300 --> 00:03:16,500 A copy of this. 94 00:03:16,500 --> 00:03:18,166 Implementation and then, you know, removing. 95 00:03:18,166 --> 00:03:20,733 All the cells and. Re-Implementing them from scratch. 96 00:03:20,733 --> 00:03:21,933 We're. Going to take instead. 97 00:03:21,933 --> 00:03:24,433 Our a priori. Implementation. 98 00:03:24,433 --> 00:03:25,033 Then we're going. 99 00:03:25,033 --> 00:03:26,533 To create a copy of. 100 00:03:26,533 --> 00:03:29,500 This implementation by clicking here on. Save a. 101 00:03:29,500 --> 00:03:32,500 Copy and drive. And then. That's where. 102 00:03:32,700 --> 00:03:33,600 You know that's on this. 103 00:03:33,600 --> 00:03:33,966 Copy that. 104 00:03:33,966 --> 00:03:35,666 I will show you how I. 105 00:03:35,666 --> 00:03:38,900 Transformed this a priori implementation into. The. 106 00:03:39,366 --> 00:03:42,000 Implementation. All right. Are you ready. 107 00:03:42,000 --> 00:03:43,200 Let's do this. 108 00:03:43,200 --> 00:03:46,633 So the first thing I did was to well, very. 109 00:03:46,633 --> 00:03:49,000 Simply change the name here. Of the. 110 00:03:49,000 --> 00:03:50,933 Ipynb file. And I called. It. 111 00:03:51,966 --> 00:03:52,366 All right. 112 00:03:52,366 --> 00:03:55,400 Let's start with the simplest change then. 113 00:03:55,400 --> 00:03:56,633 Still very simple. 114 00:03:56,633 --> 00:03:59,466 I change the title here from primary. 115 00:03:59,466 --> 00:04:02,533 To I'm really showing you everything I did. 116 00:04:02,600 --> 00:04:04,966 To make that extra implementation. 117 00:04:04,966 --> 00:04:06,166 Then I looked. 118 00:04:06,166 --> 00:04:09,166 Section after section and see if I had to change anything. 119 00:04:09,433 --> 00:04:12,533 Here we still have to install this package because. 120 00:04:12,666 --> 00:04:14,866 You know, we're building the model. 121 00:04:14,866 --> 00:04:17,366 Through a simplified. Version of. The. 122 00:04:17,366 --> 00:04:19,700 Primary model. So let's keep this. 123 00:04:19,700 --> 00:04:20,500 We can remove all. 124 00:04:20,500 --> 00:04:22,800 The outputs here because we will rerun. 125 00:04:22,800 --> 00:04:24,033 Everything. 126 00:04:24,033 --> 00:04:26,833 Then I get the three. Same libraries. Here. 127 00:04:26,833 --> 00:04:28,900 I came to the same data preprocessing phase. 128 00:04:28,900 --> 00:04:31,900 You know because indeed we still need that transactions list 129 00:04:31,966 --> 00:04:34,966 containing all the transactions into a list. 130 00:04:35,100 --> 00:04:36,000 All right. 131 00:04:36,000 --> 00:04:39,633 Then here when training the primary model on the data. Set. 132 00:04:39,833 --> 00:04:40,666 Well same. 133 00:04:40,666 --> 00:04:42,233 I kept everything we could. 134 00:04:42,233 --> 00:04:44,466 Even, you know, remove the min confidence. 135 00:04:44,466 --> 00:04:46,133 And min lived here in order to. 136 00:04:46,133 --> 00:04:47,800 Really only consider the support. 137 00:04:47,800 --> 00:04:49,766 But I recommend to still keep them 138 00:04:49,766 --> 00:04:53,200 because, you know, these two will give you even stronger associations. 139 00:04:53,200 --> 00:04:55,766 So I don't. Recommend to remove them. And then. 140 00:04:55,766 --> 00:04:57,166 I kept these because we're. 141 00:04:57,166 --> 00:04:59,833 Still in the same scenario to find the. Best deals. 142 00:04:59,833 --> 00:05:02,200 Buy one product, get another product for. Free. 143 00:05:02,200 --> 00:05:04,000 So we still have to keep this. 144 00:05:04,000 --> 00:05:06,433 But then in the end, I'll explain how to. Run. 145 00:05:06,433 --> 00:05:09,800 Some analysis on a. Larger set of product. 146 00:05:09,833 --> 00:05:11,866 Right? Because remember that. With Eeglab. 147 00:05:11,866 --> 00:05:13,333 We're not considering rules. 148 00:05:13,333 --> 00:05:15,066 But sets of. Products. 149 00:05:15,066 --> 00:05:17,733 And that's because we're only considering the supports. You know. 150 00:05:17,733 --> 00:05:22,333 The support of a set of products like let's say A, B, C, which is of course. 151 00:05:22,333 --> 00:05:24,133 The number of transactions containing the. 152 00:05:24,133 --> 00:05:26,100 Products A, B and C divided by the. 153 00:05:26,100 --> 00:05:29,100 Total number of transactions. Right. So that's why there is not this. 154 00:05:29,100 --> 00:05:32,000 Direction and. Therefore these rules okay. 155 00:05:32,000 --> 00:05:33,600 So here I kept exactly the same. 156 00:05:33,600 --> 00:05:35,166 We can, you know, change. 157 00:05:35,166 --> 00:05:37,700 A priory here by a if you want. 158 00:05:37,700 --> 00:05:39,733 Depending on how you. Want to see it. Okay. 159 00:05:39,733 --> 00:05:41,366 And then when visualizing the. 160 00:05:41,366 --> 00:05:43,966 Results that's where I'll show you what I did. 161 00:05:43,966 --> 00:05:44,900 As a main change. 162 00:05:44,900 --> 00:05:47,400 You know, as an essential change here. 163 00:05:47,400 --> 00:05:48,533 I didn't change anything. 164 00:05:48,533 --> 00:05:50,533 I still displayed all the rules. 165 00:05:50,533 --> 00:05:51,200 You know, in this. 166 00:05:51,200 --> 00:05:53,666 List of complicated structure. 167 00:05:53,666 --> 00:05:56,333 But then. Here. When, you know, putting. 168 00:05:56,333 --> 00:05:58,400 All the results, you know, all the rules well. 169 00:05:58,400 --> 00:06:00,266 Organized into a pandas dataframe. 170 00:06:00,266 --> 00:06:01,833 Well, I'm going to show. You what I did. 171 00:06:01,833 --> 00:06:03,133 Let's scroll. Down. 172 00:06:03,133 --> 00:06:05,233 Well, this time, since we no longer. 173 00:06:05,233 --> 00:06:06,700 Have. Confidences and. 174 00:06:06,700 --> 00:06:10,466 Lifts for our rules, well, very simply I took this. 175 00:06:10,733 --> 00:06:12,300 And then I removed. 176 00:06:12,300 --> 00:06:13,033 These two. Rows. 177 00:06:13,033 --> 00:06:15,300 You know, the confidences and the. Lifts. 178 00:06:15,300 --> 00:06:18,300 Removed. Of. That inspect function. 179 00:06:18,633 --> 00:06:19,500 And then same. 180 00:06:19,500 --> 00:06:22,166 Obviously we have to remove that here as well. 181 00:06:22,166 --> 00:06:22,800 Right. Because for. 182 00:06:22,800 --> 00:06:25,800 The A there is no confidence or. Lift. 183 00:06:26,066 --> 00:06:26,933 So there we go. 184 00:06:26,933 --> 00:06:29,000 And same in the columns names here. 185 00:06:29,000 --> 00:06:33,233 You know when creating the final data frame nicely visualizing the result. 186 00:06:33,233 --> 00:06:38,033 Well I removed of course confidence and lift here. 187 00:06:38,466 --> 00:06:39,266 And I even. 188 00:06:39,266 --> 00:06:41,400 Replaced. You know left hand side. 189 00:06:41,400 --> 00:06:43,266 By actually. 190 00:06:43,266 --> 00:06:46,800 Product one and right hand side by. 191 00:06:47,800 --> 00:06:48,900 Product two. 192 00:06:48,900 --> 00:06:49,533 And that's because. 193 00:06:49,533 --> 00:06:52,066 You know in the middle there is no. Rule. 194 00:06:52,066 --> 00:06:54,566 You know, we're only considering. Set of products and therefore. 195 00:06:54,566 --> 00:06:56,366 There is no question of left hand. 196 00:06:56,366 --> 00:06:58,633 Side or right hand side of a rule. All right. 197 00:06:58,633 --> 00:07:00,366 So that's what. I changed in the. 198 00:07:00,366 --> 00:07:02,433 Cell. Then. In this. Cell. 199 00:07:02,433 --> 00:07:05,400 Well I simply had. To remove it because you know. 200 00:07:05,400 --> 00:07:08,400 The principle of the model is just to. 201 00:07:08,400 --> 00:07:10,800 Return the different. Sets by. 202 00:07:10,800 --> 00:07:12,600 Descending supports, you know, from the highest. 203 00:07:12,600 --> 00:07:14,033 Support to the lowest. One. 204 00:07:14,033 --> 00:07:17,766 And therefore here we absolutely need to sort these supports directly. 205 00:07:17,900 --> 00:07:20,733 So I just remove the cell. And this one as well. 206 00:07:20,733 --> 00:07:24,300 So that we can directly display the results by. 207 00:07:24,300 --> 00:07:27,433 Descending not. Lifts but supports. 208 00:07:27,833 --> 00:07:28,633 All right. 209 00:07:28,633 --> 00:07:33,633 And of course to do this here we had to replace lift by support. 210 00:07:34,100 --> 00:07:34,700 And now it. 211 00:07:34,700 --> 00:07:35,466 Should be all good. 212 00:07:35,466 --> 00:07:38,133 Let's remove this and. Let's. 213 00:07:38,133 --> 00:07:39,933 Rerun everything. 214 00:07:39,933 --> 00:07:42,066 You know we can also remove this by the way. 215 00:07:42,066 --> 00:07:43,500 Right. We have no. Output. 216 00:07:43,500 --> 00:07:44,966 So now we're going to rerun everything. 217 00:07:44,966 --> 00:07:47,066 But first let's. Not forget to upload the. 218 00:07:47,066 --> 00:07:49,033 Data set inside a notebook. 219 00:07:49,033 --> 00:07:49,400 All right. 220 00:07:49,400 --> 00:07:53,233 So right now the notebook is connecting to a runtime to enable file browsing. 221 00:07:53,666 --> 00:07:54,800 And in a second we. 222 00:07:54,800 --> 00:07:58,000 Should see that upload. Button. There we go. 223 00:07:58,300 --> 00:08:00,500 All right let's upload. 224 00:08:00,500 --> 00:08:01,566 Then please fine. 225 00:08:01,566 --> 00:08:03,600 Your machine learning is that and. 226 00:08:03,600 --> 00:08:06,100 Data set folder. In your machine and go to. 227 00:08:06,100 --> 00:08:08,133 Part five Association Rule. Learning. 228 00:08:08,133 --> 00:08:10,866 Then section 29 Python. 229 00:08:10,866 --> 00:08:11,700 And there you go. 230 00:08:11,700 --> 00:08:15,200 Please select your data sets market basket optimization. 231 00:08:15,766 --> 00:08:16,133 All right. 232 00:08:16,133 --> 00:08:18,300 So this will. Upload it inside a. Notebook. 233 00:08:18,300 --> 00:08:19,866 In a second. 234 00:08:19,866 --> 00:08:20,566 Right. 235 00:08:20,566 --> 00:08:23,100 It's a big data set. There we go. Perfect. 236 00:08:23,100 --> 00:08:25,833 And now we're simply going to run everything and make sure. 237 00:08:25,833 --> 00:08:28,600 Everything works properly by clicking. 238 00:08:28,600 --> 00:08:29,533 Runtime here. 239 00:08:29,533 --> 00:08:30,833 And then run. 240 00:08:30,833 --> 00:08:31,766 Oh. So first. 241 00:08:31,766 --> 00:08:33,900 It will install. That a binary package. 242 00:08:33,900 --> 00:08:35,066 The same way by first 243 00:08:35,066 --> 00:08:39,100 downloading it from the link and then installing it into the notebook. 244 00:08:39,100 --> 00:08:39,600 There we go. 245 00:08:39,600 --> 00:08:42,400 Then importing the libraries and data preprocessing. Phase. 246 00:08:42,400 --> 00:08:44,733 Then the training and then the. Results. 247 00:08:44,733 --> 00:08:47,066 So here we have of course the same results as before. 248 00:08:47,066 --> 00:08:49,333 But then. For the. 249 00:08:49,333 --> 00:08:50,700 Final results, which are. 250 00:08:50,700 --> 00:08:53,566 Supposed to be the final output of the model. 251 00:08:53,566 --> 00:08:54,566 Well there you go. 252 00:08:54,566 --> 00:08:55,133 You have them. 253 00:08:55,133 --> 00:08:57,833 Here displaying the results sorted by. 254 00:08:57,833 --> 00:09:00,833 Descending support. And indeed. We see the. 255 00:09:00,966 --> 00:09:03,600 Combination of two products, you know, the. Set of two. 256 00:09:03,600 --> 00:09:04,866 Products from the highest. 257 00:09:04,866 --> 00:09:05,866 Support, oh point. 258 00:09:05,866 --> 00:09:08,233 0159, which means one point. 259 00:09:08,233 --> 00:09:10,133 Six. Percent down to. 260 00:09:10,133 --> 00:09:10,633 The lowest. 261 00:09:10,633 --> 00:09:13,400 Support, you know, for the ten highest. Support. 262 00:09:13,400 --> 00:09:16,133 Ten sets. Of products. With the ten highest. Support. 263 00:09:16,133 --> 00:09:16,966 All right. 264 00:09:16,966 --> 00:09:19,700 And that's, you know, exactly what the output. 265 00:09:19,700 --> 00:09:21,700 Of the model. Is supposed to be. 266 00:09:21,700 --> 00:09:24,100 So you see, we simply built this. 267 00:09:24,100 --> 00:09:24,666 Model by. 268 00:09:24,666 --> 00:09:27,733 Adapting. The Priory model. To the. Model. 269 00:09:27,733 --> 00:09:29,533 And returning the exact same. 270 00:09:29,533 --> 00:09:31,566 Output as it is supposed to. 271 00:09:31,566 --> 00:09:32,433 Give us, meaning. 272 00:09:32,433 --> 00:09:35,600 The set of products having the highest supports. 273 00:09:35,900 --> 00:09:37,366 And then if you want to perform an. 274 00:09:37,366 --> 00:09:39,533 Analysis with larger sets of bricks. 275 00:09:39,533 --> 00:09:42,533 Because here we only do this for sets of two products. 276 00:09:42,833 --> 00:09:44,866 Well, very simply, you just need to. 277 00:09:44,866 --> 00:09:46,533 You know, in the training. 278 00:09:46,533 --> 00:09:48,366 Cell, you just need to. 279 00:09:48,366 --> 00:09:52,100 Change these parameters from index equals to you can keep this one 280 00:09:52,100 --> 00:09:54,866 but then increasing the maximum length. And this will. 281 00:09:54,866 --> 00:09:56,833 Give you some larger set of products. 282 00:09:56,833 --> 00:09:58,700 And even if you know you will have. 283 00:09:58,700 --> 00:10:00,466 This set of several. 284 00:10:00,466 --> 00:10:03,233 Products here and then one per here, because, you know, there is still. 285 00:10:03,233 --> 00:10:05,100 This. Direction as in a rule. 286 00:10:05,100 --> 00:10:07,100 Well, that's fine, because then the support. 287 00:10:07,100 --> 00:10:08,766 Of a rule with several products. 288 00:10:08,766 --> 00:10:11,800 On the left hand side and one product on the right hand side. 289 00:10:11,800 --> 00:10:12,666 Well, is still. 290 00:10:12,666 --> 00:10:14,100 The same support. Of the. 291 00:10:14,100 --> 00:10:16,200 Set containing all. These products. All right. 292 00:10:16,200 --> 00:10:18,300 So that's how you would use this. 293 00:10:18,300 --> 00:10:19,200 Implementation for a. 294 00:10:19,200 --> 00:10:22,466 Larger set of products okay. So good. 295 00:10:22,466 --> 00:10:24,266 Let's have a quick look at the result. 296 00:10:24,266 --> 00:10:26,600 Well we have. Kind of the same ones as. Before, but. 297 00:10:26,600 --> 00:10:28,833 This time in a different. Order because we. 298 00:10:28,833 --> 00:10:30,800 Sorted them by descending. Supports. 299 00:10:30,800 --> 00:10:31,466 But there you go. 300 00:10:31,466 --> 00:10:34,033 The set of two products that. 301 00:10:34,033 --> 00:10:35,966 Appear most frequently in the store. 302 00:10:35,966 --> 00:10:39,000 You know there uppercase most frequently are herb. 303 00:10:39,000 --> 00:10:40,633 And pepper with ground beef. 304 00:10:40,633 --> 00:10:43,633 Whole wheat pasta with olive oil, pasta with scallop. 305 00:10:43,700 --> 00:10:45,533 Mushroom cream sauce with scallop. 306 00:10:45,533 --> 00:10:46,166 You know. 307 00:10:46,166 --> 00:10:48,900 All. These seem very relevant associations. 308 00:10:48,900 --> 00:10:51,900 Leading to exquisite meals cooked at home. 309 00:10:52,266 --> 00:10:52,800 Right? 310 00:10:52,800 --> 00:10:55,366 All this actually makes me kind of hungry. 311 00:10:55,366 --> 00:10:56,766 Okay, so there you go. 312 00:10:56,766 --> 00:11:00,500 So you have now an extra association rule learning model in your toolkit. 313 00:11:00,666 --> 00:11:02,400 That is all nicely. 314 00:11:02,400 --> 00:11:04,800 Adapted. From the Priory model. 315 00:11:04,800 --> 00:11:06,933 But remember my recommendation. 316 00:11:06,933 --> 00:11:08,366 I still. Recommend to. 317 00:11:08,366 --> 00:11:09,966 Work with the Priory. Model. 318 00:11:09,966 --> 00:11:12,900 Because these extra metrics, such as the. 319 00:11:12,900 --> 00:11:14,300 Confidence and the lift. 320 00:11:14,300 --> 00:11:15,000 Will give you. 321 00:11:15,000 --> 00:11:17,200 Much stronger results in the end. 322 00:11:17,200 --> 00:11:19,300 But good that you have. The two models. 323 00:11:19,300 --> 00:11:20,833 And now we're going to move on. 324 00:11:20,833 --> 00:11:23,533 To a very exciting part, which is. 325 00:11:23,533 --> 00:11:24,966 Reinforcement learning. 326 00:11:24,966 --> 00:11:26,733 And you have to know. That here we will. 327 00:11:26,733 --> 00:11:27,400 Actually make a. 328 00:11:27,400 --> 00:11:30,066 Step a lot closer to artificial. 329 00:11:30,066 --> 00:11:31,633 Intelligence. Because reinforcement. 330 00:11:31,633 --> 00:11:34,133 Learning is two branch of machine learning. 331 00:11:34,133 --> 00:11:35,900 With which you know. You can implement. 332 00:11:35,900 --> 00:11:37,800 Robotics. You know, robots. 333 00:11:37,800 --> 00:11:40,433 And of course, in part six we will not implement a robot. 334 00:11:40,433 --> 00:11:42,733 But still you will get the basics of. 335 00:11:42,733 --> 00:11:44,800 Artificial intelligence and how you can build. 336 00:11:44,800 --> 00:11:45,733 Robots. 337 00:11:45,733 --> 00:11:48,200 So I can't wait to. See you in this next part. 338 00:11:48,200 --> 00:11:50,000 You can actually hear by the sound of my voice. 339 00:11:50,000 --> 00:11:51,333 That reinforcement. 340 00:11:51,333 --> 00:11:54,666 Learning is one of my favorite branches of machine learning, and my favorite. 341 00:11:54,866 --> 00:11:57,533 Application. Of. AI. So I'll be more than happy to. 342 00:11:57,533 --> 00:12:00,400 Teach it to you. And until then, enjoy machine learning.