TG BC Study Circle | Free AI Certification Training by Telangana Govt | Day 3

追加: 2026-05-26

[00:00:00]So today we'll learn machine learning.

[00:00:04]What are types of machine learning and how it's going to work in the back?

[00:00:10]So in machine learning types of learning is supervised and supervised by supervised learning.

[00:00:23]Okay.

[00:00:27]What is then related to supervis Okay.

[00:00:57]Okay. Can you give me an example of structure data organiz Employed in Excel sheets.

[00:01:13]>> Okay. Employed in Excel sheet. Okay.

[00:01:21]What is that?

[00:01:34]And how you say?

[00:01:41]Okay.

[00:01:43]Okay.

[00:02:00]students.

[00:02:09]Every is not it's the definition of a structure but has if you want to tell what is structured data data you should you should say that structured data is unstructured data structured okay so I'm having so I insert insert.

[00:03:11]So, how you can say that?

[00:03:17]>> Yes, I did not mention any type of animal, right? So that comes under mention that this cat family and dog family mention every animal family and if I enter that structure right and sorry is known as structure.

[00:03:54]So data multiple types.

[00:04:04]So do you have types of structured structure?

[00:04:20]Okay. Relational database. Okay.

[00:04:33]Like there are main types those are regression forecasting and multiple are also comes under because NP does not have any size or does not size or only these many images we should have in the structured format right so the CV these all comes under and the navy bases and these all comes under but the structured data we going to have many types of models or many types of many types so the main type first is known as a regression right so the basic what is meant by regression What is regression?

[00:05:37]Yes. The relationship between >> two variables. Okay. What are those two variables?

[00:05:46]What ship between two variables? I'm asking what are those two variables?

[00:05:58]>> You have one dependent and you have independent.

[00:06:01]>> Okay. What is dependent? What is independent? Yes. Identifying relationship between dependent and independent variable that is known as a regression. So now you you will have an you'll have a confusion that what is mean by Yes, very good. The ice cream sales depended on season, right? If you have summer, the ice cream sales will be the high. If you have winter, the ice cream sales will be row, right? So, these are known as an dependent and independent.

[00:06:38]So, regression is nothing but identifying relationship between two variables. that is identifying relationship between dependent and independent variable that is known as an a regression right and in regression we are having multiple types linear multiple and logistic so these are the three types of variables right so do I am fast no I'm slowly right so there are three types of regression those are linear regression uh logistic and multiple regression right so what is mean by linear >> single Second B, stop talking. I want El to talk. Okay. What do you mean by linear regression?

[00:07:23]Boys, I want answers from you.

[00:07:27]>> One variable. Yes. Very good.

[00:07:29]Identifying relationship between one variable means one independent and one dependent. That is known as an linear regression. Then what about multiple regression?

[00:07:45]relationship that is known as what logistic relation is output yes and logistic if the output is in the true or false yes or no or zero or one that is known as an logistic regression right so these are three types so now uh you should say me I'll give an example so you should say me what is linear what is multiple and what is logistic regression okay so You are very uh what you say you are very excited to know my salary. Okay. So identifying of my salary is the output right. So for identifying of my salary there are three main components which are dependent on my salary. One is my age. Second one is experience and the third one is working.

[00:09:10]So there are three dependent variables.

[00:09:13]Those are uh my working hours, my age and my uh what do you say the experience. So these are three. So now see me what is linear, what is multiple and what is logistic. Compare with this example and come up.

[00:09:33]>> Okay. If I use all three variables and if I identify my salary those that is known as a multiple. Okay. Then what about linear Okay. Taking of only one variable that might be experience that might be age that might be working hard. If you take any out of three if I take only one variable and compare it with my the output that is identifying of my salary that is known as an linear regression.

[00:10:17]If I take all three variables and I if I identify one variable that is an output that is known as a multiple regression.

[00:10:24]Then what about logistic regression?

[00:10:36]All three >> Y may be right, Y may be wrong. Those are three are right. Based on that I should I identifying my salary. So identifying my salary is right and these three components also right. What is mean by logistic? How you >> working?

[00:10:57]No certain >> no I did not got if the age is like one check if the age is equal to that age then yes >> okay somewhat related yes for example if my salary is above 50,000 Right? If my salary is above 50,000 then my age criteria will be dependent on that 50,000 or not. Yes, either no. Right? So that is known as a logistic regression.

[00:11:39]For example, if you leave age criteria and if you take my experience, if I'm working in this company for one two or two or one or two years like that, if my salary is around 50,000, is this experience will be enough for this salary? Yes or no or true or false? That is known as a logistic regression.

[00:11:59]Right? So these are three type of regression. And there are this this is known as a structured data. Right? And before that one main important thing I want to say you is whatever data we are using in our daily life. So that is an structured or unstructured.

[00:12:18]>> Yes. Because we use images, we use audios, we use videos, we use text those all belongs to unstructured data.

[00:12:28]If you want to build an a website or if you want to build an a app or if you want to build an a chatboard you should get an structured data with unstructured data you cannot build those type of an website or you cannot build those type of an app. So for that first we will gather an unstructured data. After gathering unstructured data you need to convert that unstructured data into structured data. By using regression, forecasting, KN and we have multiple types in uh structured data. So by using all those models you need to convert structured data into unstructured data into structured data. after converting into structured data then you should start building of a model to get an a final output clear for example we uh has an are part of an PBL we need to create an app or we need to create an chatboard now I'm creating an app so after if I want to create an app I will collect the information regarding particular topic after collecting I'll convert this information into structured data which is an organized manner you know where the machine can understand only a structured data if I give an unstructured data to an machine so do the machine can understand unstructured data no for example let's take a chat z for example if you ask chat z to create an uh what you say animated image for example randomly uh you write the prompt that is in unorganized manner and if you click on enter do chat can understand that type of prompt which is an unorganized prompt. No. Right. The charge will ask for more verification.

[00:14:14]Right? Means do uh please me mention the situation. Please mention the uh circumstance. Please mention me the environment. So which type of people you need to add in that picture. Everything do the charge will will be asking me. So instead of that if I go to charge and clearly if I mention this type of picture I need to I need to get and this type of uh HD quality I need to get and this type of size I need to get. So if I mention all these things and if I click on enter so now charge can clearly understand my prompt and the charge give will be giving the uh image right. So that's why the structured data is more important in any other it might be in our daily life it might be any other creating the uh prompt or creating the website or any other uh coding language in anywhere the structured data is more more important clear so this is about regression and when we come to the KNN.

[00:15:16]Do you know what is mean by KNN?

[00:15:22]KNN.

[00:15:24]So, KN&N is also one type of an structured data which is going to convert unstructured data into structured data. For example, okay. So I'm having an animals data, right? So this is known as an my data data set. So now after taking animals data I want to divide all the animals data separately like rat family or like dog family like tiger family I want to divide all these families okay I want to divide and after dividing I'll take one more family that is in a cat family right so these are type of families which are from animals data right so Now what is mean by KNN? KNN means K nearest neighbor that is known as an KNN. Okay.

[00:16:34]So now under this data I I will train the machine that this is in a cat family.

[00:16:42]This is in a dog family. This is in a rat family. This is in a tiger family.

[00:16:47]I'll train the data under all these components. Okay. After training data, I'll build the model and I'm going to build the like a website. Okay. Now out of this I have been build the website.

[00:17:06]Okay. After building the website now has an a user we will uh give away this website to the user to be used. Right.

[00:17:15]After giving so you all user will enter a new data into this website. Right? So what type of data you will uh enter a dog family or in a dog picture into that website and ask uh do this dog uh for how much rate do I can sell this dog?

[00:17:35]You are going to ask to that particular website. Right? So now after entering a dog picture into this website, it will going to analyze the dog picture. After analyzing now the dog image I have been entered into the website. Right? After entering into the website, it will going to go back to this data, the previous data. Right? It will go back to the previous data and after going back to the previous data it will going to the all these type of data that it will first go near tiger and now the dog is here for example if the dog is here it will calculate the distance between this to this. Okay, after calculating the distance, it going to after calculating the distance between this new data and the old data, it going to note the distance between these two and it will note somewhere. After that, it will move to here. After moving here, it will it is going to calculate the distance again from this to this.

[00:19:02]After calculating the distance between this to this, it going to note the distance between these two components and after that it will go near rat and after that it go near cat. After noting all this distance uh how much distance it is there from new data to the old data. It going to note as a 1 2 3 4. We have four uh old data. Right? After noting the old data which is the nearest form means now this is the somewhat distance and this is somewhat near and if the from here if if it is going to calculate the distance and from here if it is going to calculate the distance this is the more and this is somewhat less and from here if it's going to calculate the distance it is also more the distance between this to this is less right so now by using K and Then it going to assign this new data into this old data means now it going to calculate that okay this is known as an a dog. So now this is this new data will be going to add in the old data that is in a dog family. After adding now by using AI it going to give us the answer that how much this dog is going to be considered.

[00:20:19]Right? For example, even in in your Google photos, your features might be matching with the other people features also, right? So that's why sometimes the other people picture might be there in your folder automatically, right? How it going to happen means the machine is going to analyze in the wrong direction or else it identify that your phrase structure and the other people face structure might be same or similar. That's why that people photo will be in your folder. If you click on wrong and if you say that machine th this picture is not me this picture is the other person. So from next next time this picture will be not entered into your uh into into your Google photos. Right? So this is how the K andN is going to working. Clear? The K nearest neighbor means how much the nearest data will be there. the shortest nearest data to that nearest data the new data will be entering into the old data. So based on our prompt now the AI will going to give us the answer. So that is known as an K nearest neighbor.

[00:21:28]Clear how the K nearest neighbor is going to working yes no any doubts how it's going to work and this is about the KNN. So we'll move forward for the forecasting right. So what do you mean by forecasting?

[00:21:48]I know you should answer me because the forecasting is very easy.

[00:21:53]>> Prediction. Okay. What do you predict?

[00:21:57]>> Okay. Based on which do you predict the weather?

[00:22:03]>> Previous previous year. Okay.

[00:22:08]Convert into sentence format. take the previous days.

[00:22:17]Yes, it takes the uh future years prediction or future year data which is have been recorded from that it going to take it and by using the forecasting it's going to predict the uh future. So I I have been past data it's going to take the past data which already has been recorded in the last year. For example, if we take the uh May month, how much degree Celsius do the last year we got the in in May month, if you take that and if you if you want to predict this May month uh degree. So based on that time series, it's going to predict the data. So main thing you need to say in the forecasting is time series.

[00:23:07]So time series is the most important part in the forecasting itself. So based on your time you are going to predict the future or you are going to predict the past or you are going to predict the present that is known as an forecasting that might be weather that might be any other forecasting that depends on a time series basically depends on a time series. Clear the forecasting and we have so many types in the structure but those are nothing more important because has if you want to know about more about coding those are very important but as of now this much of uh structured data is enough and uh when we move for the unstructured data as I told NLP belongs to unstructured data and CV belongs to unstructured data the Navy B bias is also belongs to unstructured data.

[00:23:59]and the text, images and the audios, videos all these comes under unstructured data. So whatever we are using nowadays all types of data belongs to uh the unstructured data. Clear? And this is about the structured and unstructured. So we'll move forward for the PBL. Before that uh I will say how to create it or how to arrange the your PBL format and everything. So for PBL and this PBL need to be done by today itself right so by sitting in the class you need to create whatever your website or your app or anything whatever you want to create it clear so for PBL it will work.

[00:25:01]So yesterday you have been identify your problem and your mentor means on Friday right you have been identify your problem and you have been identify your solution right. So now after identifying your problem and after identifying your third step is okay the third step is implementation.

[00:25:35]So now you should implement your problem and solution into some AI or AI platform right so have you having your paper that is which have been written in the paper your problem and your solution only the one person have been g given me I think who I don't remember from girls you have been given me okay so now take your problem statement and solution right now you should come up with what you want to create it that you want to create a website or you want to create an a chatboard or you want to create an app right now you should go to chat JPT and ask chat JPT to give a best prompt because we are not in a prompt engineerings right so we cannot uh give a best prompt to that AI platform right so go to chajipity and ask charge to give a best prompt so that you can create create your website or you can you can create your app or you can create your chatboard in that particular site right for example you are having multiple sites I'm I'll send the site names in your group so you can watch your group that the most there are two useful websites those are vixwipe and lovable lovable so lovable.ai AI vix wipe are two most used platforms to create chatboard or to create an website.

[00:27:05]Clear? So now after going to charge for example my problem is stress management.

[00:27:11]For example I'll go to charge and I'll ask charge to give me a best prompt based on stress management so that I can create a chatboard on the stress management topic. I'll go to charge and I'll ask charge give me a prompt my problem statement is stress management I want to create an app or I want to create an website on lovable so based on that give me the prompt so after giving this prompt after clicking enter charge is going to give you the prompt after that copy that prompt go to Google and type lovable or vix wipe and after typing you'll get in a first site that is in a lovable or in a vix swipe I'll show in the project I'll show in laptop after getting after clicking on that particular site sign in first okay after signing whatever prompt you haven't pasted copied from shipy paste that prompt in lovable after pasting click on enter leave it for 5 minutes or 10 minutes after leaving it's going to create a website right once I have on sat on Friday I have been shown one website right for example you have four tabs home tips and quiz and pledge when I'm clicking on home homepage is v visible when I'm clicking on tips tips page page page is visible and when I clicked on quiz quiz page is visible and I can do the quiz also so like that your website need to be there it not to be recorded at the back end but at front end it should be showing me how it's going going to be happening or how it's going to working for example chatbot also if I am typing any type of question in chatboard the chatbot need to be giving answer to me right so like that you need to create it once after creating if you're not satisfied again go back to chad GP and ask charge this is not satisfied prompt so I want this this type of edit so give uh explain to charge and now again ask charge to give me the new prompt. So after giving again copy that prompt and paste that prompt in the lovable again it it is going to create your website clear. So for one email id you can create like three or four times now than that you can you canot create it. So if you have other email id means the other one email id so you can use other email id and you can again login and you can continue with the creating of your website. Clear? So start doing it. I'll show you how to come do it also.

[00:30:23]You may ask me ma'am uh what is the use of this workshop if I ask chargity of giving of a prompt right? So that is where within 5 days you cannot uh get an idea full idea how to generate a prompt or how to give a prompt right. So while you're asking charge JP to give a prompt you should observe how charge is going to uh what you say it's going to give you the prompt. So that you need to note and you need to implement in the other times the prompt giving or any other thing right. So okay prompt engineering that is going that is not going to work for prompt engineering outside they're playing lags of money for prompt engineering role right. So within few days if you want to create an best brand that is not possible that's why we are taking charge help not we are totally depending on charge we are taking charge help and we are creating a website right first put all your thoughts in charge I want to create this type of website I want to have this type of icons the structure of website need to be like this the color need to be like this and this type of images need to be added in my website give everything in prompt and Ask charge to create a sequenced prompt so that the charge will be creating a sequenced prompt. Right? Clear?

[00:31:46]Meanwhile, I'm going to show you how to create And today last half an hour before you leaving the as we'll also create our GitHub profiles and after creating tomorrow you're going to give the presentation on your website whatever you are creating now clear how many tabs you are having or what type of explanation you'll give that is your wish but tomorrow you need to come front and you're going to give the presentation on your website clear if your team members are absent how many are coming for tomorrow all those are coming front and you're going to give the presentation Clear boys mainly you should give compulsory you should confident and you should give definitely your presentation clear I cannot understand I I can understand only three languages dead Amen.

[00:34:09]You guys don't Excuse me.

[00:35:59]Yes, you can see on the projector.

[00:40:00]on what connects.

[00:40:58]How about this?

[00:41:31]Oh, nice.

[00:42:25]Yes. So guys once see to the projector.

[00:42:28]So for example uh I'm going to charge a pity.

[00:42:43]So we should off the lights or no need.

[00:42:46]Boys can you visible off the lights?

[00:42:57]See that?

[00:43:48]Here we go.

[00:45:09]See on the projector I want to create a website on lovable. So my topic is stress management. So now give me a prompt.

[00:45:27]Okay. And I'm clicking on enter.

[00:45:34]So you can see here how the charge is giving me the prompt. The website name, the main goal and all these I need to copy.

[00:45:49]Okay. From here I'm from here I'm going to copy the website name the mind easy.

[00:45:56]So I'm copying it.

[00:46:01]copying it and I'm going back to Google and here I have been typed lovable. So this is the first one that is the website. So I'm clicking on this lovable app and now after clicking so this is the interface you'll get for the lovable.

[00:46:19]Okay. And after clicking you should log in first with your email id.

[00:46:28]So whatever your email id you want to choose just choose your email id.

[00:46:41]So already I have account for this email id. So directly it will be logging to the to my lovable site. So for you you did not have your account in lovable. So it is going to ask for your few of the permission like solo or you want to do it in your team work or for which you want to create this lovable every question will be asked. So fill all the all the questions and after filling you'll get the format like this. Okay.

[00:47:09]Now whatever prompt I have been copied from chat zpity I'm going to paste that prompt here. See ask lovable to create a blog about. So I'm going to paste that prompt here. Right. After pasting the prompt here, I'm going to click this up arrow. Okay, after clicking okay here you can see that your app or your website your chatboard is going to created here. So I did not mention I have already said chat GP that I want to create a website. So based on that give me the prompt. So if you want to create a chat board so ask chat GPT that I want to create a chat board. So based on that give me the prompt. So like that the chat GP is going to give the give you the prompt. So copy that paste that prompt here and if you click on login the same way login with your Google accounts.