Speaker 1: There are 500 million tweets per day and 800 million monthly active users on Instagram, while 90% of whom are younger than 35. Users make 2.8 million Reddit comments per day. So there are huge amount of data we generate every day. And it is getting extremely difficult to get the relevant inside out from this clutter. So how can you actually use those data? So in that case, sentiment analysis comes into the picture. So sentiment analysis is a subfield of natural language processing that tries to identify and extract the opinions from the review of respective blogs. So if you want to see from some of the feedback that the feedback is stand for positive feedback or a negative feedback or a neutral feedback, that time we use a technique that is called sentiment analysis. From where we are getting this feedback, we are getting this feedback from a guest information, maybe the portal management software or restaurant software or feedback front. So there are multiple resources from there. We are getting this feedback. So now have a look at what are the types of sentiment analysis we have. So we have emotion detection. We have aspect based sentiment analysis. We have fine grained sentiment analysis, and we also have multilingual sentiment analysis. So what is fine grained sentiment analysis? So fine grained sentiment analysis is basically done at text and sentence level. Metas like steaming, bag of words, bigram, tigram and sentence level features are used to understand the sentiment polarity. Maybe your feedback can be very positive, normal, positive, or can be neutral, negative or very negative. So these are the domain or these are the category we have. So when we are going to check the sentiment for a particular review, now, let's have a look at what is a state based sentiment analyst. So this breakdown stakes into a state that is attributes or components of a product or service, and they allocate each one are sentiment label. Maybe positive, negative, or neutral. So this is how our aspect based sentiment analysis works. Now, let's have a look at what is emotion detection from the name itself. You can understand. We are going to speech emotion by looking at the. person. So how can you do that? So emotion detection is the process of identifying human emotion. People vary widely in their accuracy at recognizing the emotions of others. Use of technology to help people with emotion recognition is a relatively nascent research area. Works best if it uses multiple modalities in context. Work has been conducted on automating the recognition of facial expression from video, spoken expression from audio and written expression from the text. So this is the way you can detect the emotion for a particular person. So you can see in this person how a probability of anger have a high probability of neutral sadness and all right. From this, you can see maybe the person is neutral because you can see this is the number we are getting over here. They are the probabilities. So looking at the probability is actually very useful for identifying. From it. I could identify which person has the probability is a zero point probabilities for probability high is for neutral. So you can see this particular person emotion is neutral. So in the second case also, you can see the neutral probability is 0.9. So in this case, we can quite sure that this person emotion is neutral. So this is how we actually go for detecting the emotion for that particular person. We'll have a look at other ways of doing that. Maybe here you are getting another applications of angry, happy, sad, surprise. So it's up to you. What are the emotions you want to add to your application, right? So these are the main, or you can say these are the one of the best use of sentiment analysis. Now we will have a look at multilingual sentiment analysis, but what is multilingual sentiment analysis? The sentiment analysis done in multiple languages done by the use of complex neural networks architecture. Like RNN, LSTN, many pretend models are there. So pretend models are basically the models who are trained with a huge number of datasets. So in that case, when you are, you are going to use a pretend model, you don't need to write a model from a scratch. What do you need to do? You can just change the output layer according to your users. So most popular trained models, we have Google's part and Excel net, Excel net two as well. So these are the pretend model option. You have. If you want to make a pretend sentiment analysis model, and that will not take a long time to do that as well. Now we are going to talk about neural network, so we can do the sentiment analysis using machine learning as well. But if you want to make the sentiment analysis in an advanced way, that time you need to go for neural network. But what is that neural network stands for? Basically, neural network is like a replica of human being, but the scientists try to make a replica of human being, which can detect and walk like a human brain, right? So there are lots of algorithms and lots of parts we have in neural network who can do the sentiment analysis like a human. Now we will have a look at RNN model. So basically RNN stands for recurrent neural network. So we use recurrent neural network model to do the sentiment analysis. And this model is quite familiar to walk with the sequential model as well, but why I'm talking about sequential model? Because the sequential model does maintain the insertion order. So whenever you are going to make a sentiment analysis for a particular sentence, you need to maintain the sequence as well. So this is how RNN model is recommended for sentiment analysis case. Now we have lots of other algorithms as well, like LSTM that is more better than RNN and we have GRU as well. Now we will have a look at what are the companies that are currently using sentiment analysis. Already I have given you the idea that what are the companies already using sentiment analysis. I think almost all the companies are currently using their data to understand their product quality and to understand the customer sentiment using their data. The companies like Amazon, Flipkart, Google, Facebook, Zomato are currently using sentiment analysis. So we almost came to the end of the short video of what sentiment analysis and how sentiment analysis works. And what are the companies currently using sentiment analysis? So we have a lot of companies that are currently using sentiment analysis. So if you want to learn more about this technical content, then please subscribe to Great Learning and like the video. And if you want to do courses on this technique, then please go to Great Learning Academy, where you will get almost 80 plus free courses. And after completing your course, you can claim your certificate. And if you want to do this process in your mobile application, that is also possible using Great Learning app. Thank you so much. Bye-bye. See you next time. Thank you.