Okay, so in this chapter we are going to talk about the the SDK test Explorer. Okay. So, in this chapter we are going to cover the SDK test explorer settings. Understanding the the SSP application, we test Explorer, modeling evaluation deployment with test Explorer. Okay. So let's do Eddie.

Ss para. So, Gabby, who is SS Bora you can only use this software for seven times more Okay, so this is the SDK test Explorer. So I'm going to set settings just like the SDK script, right? The SDK studio. So I will have to set the DST Okay, the die of the sea. Fun and then the SDK script writer and I agree, okay.

And other Data Explorer Okay, you can only use the software for say time says time small. Okay, so I'm going to import some tests the time to these tests. So let's look at what kind of data we are going to import. Okay, so this is the data I'm going to import. Okay, I have this is rather data size Okay, so this is where the data set can be downloaded. So for myself I have a strategy data.

I have a strategy data and I play and Amazon sells labor on TST okay so I'm going to employees for a drive thus far amazing cells they bought yesterday again is like a CSV file. So I preview and then okay so you are asked me which of robot sista which are the robot contains all the documents. So documents is all the charge So, I will Salah cumin and then okay. So, as you can see for sure data every row is a testing pool. So, in this one every row is respond from the user okay and his role is Taya ishod ma all in text form. So, this is where we will be using text mining or data analysis when when the data do not have any numeric will not have any numbers for you to process or analyze.

We have to use text mining or data analysis. Okay. So if you want to view the document in a shop law you can just double click, double click on the on the role and the document will change here okay. So you will also see some highlight. So this rate Allah is actually negative sentiments. Green color is positive positive sentiments.

Okay these are this Rick recolor green color order sentiments you can find that dictionary inside here. pasa TV here are these works. Okay, that TV audience works Okay. So you okay in your own words okay and I think you can delete also. So this is it here is a T is based on the gate kick is a this is an open source software for the data analysis also. So I'm using the gate is a tear here okay if you click on any of the cases here he may try to highlight or worse okay based on all these misspelling inside these here okay you can change the color change the color of the hog is a test highlight Change the color of positive sentiments, change the colors of negative sentiments.

Okay, of course you can also at the end and in your own words into the KCTS. So this is the same analysis. Okay. So far the understanding stage, you can go to the tools and then do all these data analyses. So, let's say for now I'm doing the POS again, which is part of speech tagging. Okay, so you can see the document here is the POS tagging.

So this is should be pronoun this one should be some book and and this one kid This one is a now this one is a book and then this one is above and this one is attractive and so on. So this is POS tagging part of speech tagging Everything going to name and ID. So I can be music group of authors okay can be a country code okay did mean a dog can be a determiner and it can also be a country code. Line life can be organization downs and so on. So this is a name and type the tagging. Then we have sentiment analysis So far this command here we have a tree was positive, who was D five was a strong one way is weak it is a passive one pleasure no pain no feel no arousal no emotional virtue no bias or no okay.

So, these are the sentiment analysis. Then we have the average TF IDF So, this is the temporary currency and inverse document okay. So, this one is basically to use to find the importance of the water Insert order documents. So we have these you can put into Microsoft side and slot them and then see which of the words are more important. And then you do something with ah da okay then we have teslin analysis Okay what testing analysis do is that you will try to get order now in every document type now and near was active and then show something like this case an excellent then case go value excellent conversations more the swan has not made much sense to me major more ozone as a mass sense problem or okay bra right volume the sun those are silver those and silver okay contestable fans and money needless so what sin analysis is trying to do is to Ah, show us all try to summarize all the documents and try to show us the important information.

Okay, this is actually part of the SPSS Statistics analytics. Okay something like this one Okay subbies was always useless okay Then contrast to be okay for sub z. So this is something this that's it. This is excellent analysis in SPSS the end SDK also of testing and analysis are using a in the now in the document and get an error surjective. So we will have something like case is excellent cases good. So we've slid into races, we can pre create a view and understand what what all the documents are trying to say, say about something.

So he gave us a summary of the order documents in the data gateway Ah, generate up ID, the SDK that has power. You can then use a software like Jeff gay to draw social network analysis graph from the data to represent something like a SPSS testing analysis, chop you can use this software to generate something like this. Okay, social network analysis. So each of the what is called the VA is tada Nope. Okay then guys, Li H or ln connecting between these two words. Okay.

So one now one adjective If connected together okay so you can after generating this data, you can put it into Microsoft Excel, how many? How many times this link as is and then put it into software ID he took out the destiny analysis gave it you can roughly know why is inside the documents you have a summary of why these documents were order documents trying to say okay in the previous are the SDK to version I do use Python two For all the social network analysis crap by the Python I use NLT, Kay and NLT Kay digitize a few key few gigabytes. So I will write this destiny and see the the SDK tree. I rewrite them allow the machine learning algorithm ID, K and Rhonda I write them myself by using all those libraries. So the current version of the SDK tree the file size is reduced.

Ah, by these features like this testing analysis, I will also need some some time to write the social network, social network graph or social network analysis graph. So far, the current version you can Take this data put it into Microsoft Excel how many times Celine PS Okay. And then you put into sabella Jackie to proudly so a destiny analysis chart okay. So we i d i have has broken news the data understanding we bodies are tagging it sentiment Elysees, English TF IDF teslin analysis. I can roughly understand or visualize the all the documents in this data set so, I Rather understand the data. So, the next stage will be proceeding to the data pre processing or the data preparation stage.

So, for data preparation test Bara allows us to convert order to Q main to upper case lower case proper case. So, I can convert this document into upper case. So, but I knew a robot name NyQuil key okay upper case and I also okay. Ah do some stemming or pata stemming on the document so, Okay, so you see a value the ease remove. So this is a POTUS me I can also remove those words okay so you can compare data Okay so after removing all the Star Wars he becomes very proud in unless competitor okay. So for this I will find as removed.

So, these are the three processing and we can remove the remove a regular expression or so. Okay and maybe I would like to use slash s plus k near me see the man I eg okay remove all that spacey Okay and then they can convert a document to test vector okay? And so on. So, for this document you see I keep using an ISA available, you can actually change the comparable to some other variable name. Okay, so let's say when you process the, ah Don't be a robot and convert them to lowercase uppercase, and you can select this variable here and then from this.up book, no comma uppercase, then you do another processing and so on. So with that, I really cover the data preparation stage.

So now we are going to watch the modeling and evaluation and deployment stage. So this is classification. Okay. So far this chapter, the test classification. Okay, I'm going to use I'm also using the sentiment, labor data from the UCI as I shown just now. So I'm strike RDS data for better or worse string dot CSV training training file and then all this data for the testing file that our was our test dot CSV so I'm going to imply the data tree and the battle was tested.

And any iOS OLED best, best or worst algorithm. So the column should be Based on what training data okay documents should be zero and a degree should be one saw the saw the input as variable should be the document. Y variable should be the category okay and then if the way okay so the accuracy is 100% then for the same thing I can use it to predict Okay so that's all was Andy algorri Danny's backup works should be zero Why should be one and the MPD. Okay, generator is better word strain dot CSV predictor dot CSV. So I'm going to D drive and I'm going to put in here what is a CME Okay, I'm going to throw in here okay 010 Okay. So these are the bodily evaluation and the deployment stage.

So let's say I want to use other algorithm. So I have to compare my test to vector. Okay. Okay, so this is a test. So you can see I convert up from, test your data and then convert them into numeric data So from here you can modify this data into the training data and testing data. Okay.

So far here has a tool for go rhythms and naive Bayes. They may not set any numbers more than one. So you'll have to do some normalization and then pointing to the algorithm into the algorithms for the modeling and the evaluation. So with that, I Ah, In that we have come to the end of the chapter phi, which is Maundy using the SDK tax Explorer. Okay, in this chapter, we have explored the the SDK as a smaller settings, as usual, like the SDK studio is the CMT. We have do the data, understanding the Data Explorer.

Data Preparation is that explorer and the modeling evaluation and deployment with the test Explorer. Okay, so in conclusion, we have completed Audi Audi costs. So we have run through the script, right? Essentially, we have run through the SDK studio, we have run through the SDK. fora, Toby that Gula and have a good journey here, just a tiny test my name using the SDK tree software. And we are we have come to the end of these costs.

Gay. Thank you very much for your participation.