Hello everyone.
This week I would like to put under our discussion
the Sentiment Analysis.
People around the world are more
actively using social media platform such as Twitter, Facebook, and Instagram
etc. They share information, opinions, ideas, experiences and other details in
the social media. The analysis of brand specific opinions can inform companies
on the level of satisfaction within consumers.
The sentiment analysis can be used to
estimate the emotions of the users objectively.
At its core, sentiment analysis, also
called opinion mining or emotion AI.
Here's
the link to the articles:
https://www.researchgate.net/publication/264840229_Sentiment_Analysis_An_Overview
My questions:
- Do you think is possible for an algorithm to analyzed meaning of the text as accurately as human?
- What are the applications of sentiment analysis?
- Why and How Companies Should Use Sentiment Analysis?
1. Do you think is possible for an algorithm to analyzed meaning of the text as accurately as human?
ReplyDeleteFrom the technical side, I think that the algorithm can even more accurately analyze it. However, when it comes to content analysis, I think that a better analysis will be done by a human being. If we take a standard example: Adam Mickiewicz Dziady part III and number 44. Which algorithm will analyze and say what the author meant? ;-)
2. What are the applications of sentiment analysis?
I have no idea! Maybe in telemarketing? When a man begins to talk about something, emotions emanate from him. I have never met such software so I have little ideas.
3. Why and How Companies Should Use Sentiment Analysis?
Maybe I will answer it straight, but companies should analyze the emotions associated with their products. When the new washing machine breaks down for the nth time by calling the service and ordering the repair service, we usually emotionally refer to the service technician. Such an analysis could allow to determine which products the customers are really satisfied with.
Hi Michal, thank you for your comment. Analyze poetry or old-fashioned texts is in definitely more complex than analyzing social media messages. The applications for sentiment analysis are endless. More and more we’re seeing it used in social media monitoring and VOC to track customer reviews, survey responses, competitors, etc.
Delete1. Do you think is possible for an algorithm to analyze meaning of the text as accurately as human?
ReplyDeleteI think that sooner or later it will be possible, in the end it is a dictionary knowledge and grammar, which can be well written and create a translator with human skills, and even better. There are many more complicated algorithms, so I think it is absolutely possible.
2. What are the applications of sentiment analysis?
The applications of sentiment analysis in business cannot be overlooked. Sentiment analysis in business can prove a major breakthrough for the complete brand revitalization. The key to running a successful business with the sentiments data is the ability to exploit the unstructured data for actionable insights.
3. Why and How Companies Should Use Sentiment Analysis?
Having insights-rich information eliminates the guesswork and execution of timely decisions. With the sentiment data about your established and the new products, it’s easier to estimate your customer retention rate. Based on the reviews generated through sentiment analysis in business, you can always adjust to the present market situation and satisfy your customers in a better way. Overall, you can make immediate decisions with automated insights. Business intelligence is all about staying dynamic throughout. Having the sentiments data gives you that liberty. If you develop a big idea, you can test it before bringing life to it.
Hello Cezary, thank you for your thoughts. As you remark sentiment analysis is in demand because of its efficiency. Thousands of text documents can be processed for sentiment (and other features including named entities, topics, themes, etc.) in seconds, compared to the hours it would take a team of people to manually complete.
ReplyDeleteDo you think is possible for an algorithm to analyzed meaning of the text as accurately as human?
ReplyDeleteI think that it dependes on the main topic of analized text. Reviews of restaurants, products or movies are easy to understand for both humans and algorithms. When it comes to sentiment analysis of news, tweets etc it may be not so easy. We have to have some background knowlegde to understand hidden meaning like irony. It is not easy for humans and even more difficult for machines but noone proved that it is impossible.
What are the applications of sentiment analysis?
I think that the most imortant thing is application of sentiment analysis on huge amount of texts so it helps to analyse it faster than humans. It can be applied everywhere where the opinion is expressed but not only. There is also subfield of fact-based sentiment analysis.
Why and How Companies Should Use Sentiment Analysis?
They can use sentiment analysis to check how their products are rated, what are ipinions about them and their competitors. They ay analysis which aspects of their products are good or which they should improve.
Hi Katarzyna, thanks for your comment.
DeleteSentiment analysis is also used for feedback monitoring, it can provide a filter about, which service of the company is getting more negative feedback. This will help the company to know, what are the problems arising with that particular service. And based on this information the company can rectify these problems.
1. Do you think is possible for an algorithm to analyze meaning of the text as accurately as a human?
ReplyDeleteIMO, there is no definitive answer to this question. Not because of the lack of computing power, but because people often misunderstand eachother anyway. The nerdy science and tech people often give up on answering the cardinal question in Polish courses: "what did the poet have on his mind?".
2. What are the applications of sentiment analysis?
I see three main customer groups for this kind of service - politicians: to assess responses to their policies and predict their election chances at the moment (e.g. when considering a premature election), marketing people - to see what people like, and how was a marketing strategy received, and finally - so called 'moderators', i.e. people busy with censorship. It may help them ban and deplatform other people, whose tone they don't like.
3. Why and How Companies Should Use Sentiment Analysis?
Let's first get the 'why' out of the way, because it's not a new phenomenon. Companies have always analyzed sentiment, but it was enough to read the newspapers, talk to lower management and find out about the public image of the company. The difference today is, there is just so much more information, opinion and also sentiment, that it can't be simply processed the old way. So computers can help, but they shouldn't be used as a black-box magic oracle, but as an efficient tool to funnel and filter the enormous amount of information produced every day.
Hello Tomasz, thank you for sharing.
DeleteI didn't hear about using sentiment analysis to support blogs moderation, but i think it's great idea. I could speed up the process of blocking inappropriate contents.
1.Do you think is possible for an algorithm to analyzed meaning of the text as accurately as human?
ReplyDeleteI think that the algorithm analyzing the meaning of the text might be more accurate than the human but the context is the most important in analyzing text and the automatic algorithm could have a problem with it. There are cases that context can change everything. On the other hand, the evaluation of human will be subjective.
2. What are the applications of sentiment analysis?
I found some applications of sentiment analysis. The sentiment analysis gives business valuable insights into how people feel about product brand or service. It can be used to identify when potential negative threads are emerging. Sentiment analysis could also be applied to corporate network, for example, by applying it to your email server. Emails could be monitored for their general “tone”. For example, Tone Detector is an Outlook Add-in that determines the “tone” of your email as you type like an emotional spell checker for all of your outgoing email.
The sentimental analysis are also applied in business intelligence. Here is the link to the article: https://www.researchgate.net/publication/290595280_Sentiment_analysis_in_business_intelligence_A_survey
3.Why and How Companies Should Use Sentiment Analysis?
Sentiment analysis has many applications and benefits to any business and organization. As I mentioned above, when the negative threads about your company appear in the Internet, you can be proactive in dealing with it more quickly. The company can obtains the feedback about their products and activity. People in some companies write a lot of emails, some of them are very important so company can use e.g. the Tone Detector to define the tone of e-mial in order to not offend someone. The companies might use the sentiment analysis to define the level of client satisfaction as well.
Hi Monika, thank you for for your comment.
DeleteAs you noticed company can use sentiment analysis in many ways, another one is get to know what’s trending. This will not only help the company to stay updated and connect more with the audience, but it will also facilitate the rise of new ideas, for developing new products. This will allow the company determine what the majority of the audience demands, and develop a product according to these demands.
Hello Klara,
ReplyDeleteThank you for a fascinating subject and links for the reading. The emotion AI is a fascinating subject indeed. I think that it is tough for an algorithm to analyse the meaning of the text as accurately as human, at least- in particular cases, because very often there is the second layer of our intention (e.g. sarcasm).
However, I think that the applications of sentiment analysis are far-reaching and they could be applied in the diagnosis of mental state. Let's consider depression for example. Social media/blog posts monitioring could emerge an alert that something bad is going on with the author. That could lead to mental health monitoring online program, which is especially crucial in the country, where almost half of the population suffers from the depression. Worth to mention that this disease is also strongly correlated with neurodegenerative diseases too.
From the other hand, I see great commercial potential for the application of sentiment analysis. First thought: social media monitoring of the mentions about the brand. This kind of algorithms would be (sometimes: already "are") an excellent improvement for the tools like Brand24 or SentiOne.
Cheers!
Hello Artur
DeleteThank you for this great example of sentiment analysis application outside of business field! Patients with depression was neglected by society for a long time, so I think that is great that we can use AI algorithm to look up for people that need help, and we can do it for broad scale.
Do you think is possible for an algorithm to analyzed meaning of the text as accurately as human?
ReplyDeleteI dealt with this topic, but with the use of statistical teaching methods on which the previous methods of machine learning are based, it does not make much sense because it requires the calculation of N-grams. The use of n-grams requires the collection of correspondingly large amount of statistical data - the corpus. The creation of an n-gram model begins with counting the occurrences of a sequence of a fixed length n in the existing language resources. Usually whole texts are analyzed and all single occurrences (1-grams, unigrams), two (2-grams, bigrams) and three (3-grams, trigrams) are counted. In order to obtain 4-grams of words you need a lot of language data, which is difficult to achieve especially for the Polish language. And to apply the context of the trumpet use an infinitely large dictionary. For the Polish language the analysis of texts containing 300 000 000 words gives a good 1-gram model and a satisfactory 2-gram model. In order to create a reliable 3-gram model, much more resources are needed. A large amount of analyzed text improves the quality of the model, although there are also methods to improve the n-gram models without additional data, based on smoothing the collected statistics. This is not possible using the existing methods.
What are the applications of sentiment analysis?
Why and How Companies Should Use Sentiment Analysis?
The most needed will be in order to autmatically translate the text.The phraseological context for most translators is doubly troublesome. First of all, because we are not able to know and remember the meanings of all idioms. It also happens that many of them do not appear in popular dictionaries or encyclopaedias. Giving up the sense of such a translation becomes a problem. When searching for appropriate sources of phraseologies, we should first of all pay attention to the elements of etymology of a particular word or expression. This is often the best advice for further action. On the Internet there are encyclopedias of myths and legends from all over the world. It is worth to get used to them. They can help us recognize characters that do not appear in Polish phraseologies
Hi Andrzej,
DeleteThank you for your comment, I'm very interested in your opinion as a user of n-gram model for sentiment analysis. What was the use case of your work? Did you think about redoing it with contextual semantic search, I think it may give you better results
1. Do you think is possible for an algorithm to analyzed meaning of the text as accurately as human?
ReplyDeleteI think that these algorithms can only approximate how people perceive sentiments, but they will not be able to do it as well as we do. There are many variables here and even people are not always able to determine what the author meant.
2. What are the applications of sentiment analysis?
The spectrum of use of such analyses is very wide. In fact, it can be used wherever we deal with text. As someone mentioned, we can analyze blog posts, comments, opinions or social media posts and it can help in moderation but also discover what kind of social moods are.
3. Why and How Companies Should Use Sentiment Analysis?
Companies can use this analysis to see how people react to their products. I think they can also help with business analysis and decision making. They will be able to make decisions based on what the customer community wants, and that is always good for both parties.
Hi Przemysław, thank you or shearing.
DeleteAs you remarked sentiment analysis can by used to amplify voice of customer. Social media reviews, mentions and surveys help to broadcast the voice of customers to the brand, they are expressing their views about. This way the brand knows, exactly how common folk feels about their services. The company can use this information in growing their market, advertisement targeting and building loyalty among its customers.
1. Do you think is possible for an algorithm to analyzed meaning of the text as accurately as human?
ReplyDeleteMaybe in a future, currently it is more like a statistic, so each individual text is not important, only the general meaning of all text metters. But when we starting constructing chat bots, then accurate meaning of every sentence is important and new chat bots already started this path.
2. What are the applications of sentiment analysis?
I was involved in one application of sentiment analysis for NASCAR. During each race algoryth analysied the whole flow of conversations in FB or Twitter and transmission operator saw what is hot topic now, what people like and if i.e. people start taking about same accident he knows, that they want to see more details of this situation and he start to show them more replays.
3. Why and How Companies Should Use Sentiment Analysis?
There is a lot of different applications, i.e. It is really important to quickly react on every image problem, so based on Sentiment Analysis companies should quickly correct all PR mistakes.
Tomasz thank you for your thoughts,
DeleteBeing involve in project of that scale must by great challenge.
Can you share with us what was the amount of twits / FB comments processed by the algorithm in total, or per minute?
Do you think is possible for an algorithm to analyzed meaning of the text as accurately as human?
ReplyDeleteYes this is possible but reasoning or true general artificial intelligence is way ahead of us and a lot of research need to be done.
What are the applications of sentiment analysis?
IT could help with reasoning, product recommendation, aggression detection and fake news detection.
Why and How Companies Should Use Sentiment Analysis?
E-commerce companies are actively using sentiment analysis for product recommendation same for music / video streaming companies. All of above companies are using for cost optimisation and as well serving better products for end user.
Hello Ewelina, thank you for your comment.
DeleteI agree that we have a ton of work to do before we will be able to manage a full potential of AI. I think that this progress is done day by day and it's possible thanks to applying AI based algorithms just like the sentiment analysis to every day solutions.
Currently algorithms can't analyse meaning of texts as humans do. There is no algorithm that comprehend a text. But algorithms may solve some specific problems in a specific setting and do it with better accuracy than an average human.
ReplyDeleteSentiment analysis can be used to for example analyze quantitatively what kind of content meets a positive reception from users, analyze comments to detect if they contain hate speech etc.
Companies can monitor comments related to their products to improve them.
Hi Adam, thank you for sharing.
DeleteThe sentiment analysis is used by the companies in many ways, as you spotted - customer feedback is one of it , other used are: finding hot keywords, getting to know what’s trending or improving services.
1. Do you think is possible for an algorithm to analyzed meaning of the text as accurately as human?
ReplyDeleteI think its possible for simple texts and its depends on language. The algorithms are evolving fast and maybe in the nearest future it will handle witch any text.
2. What are the applications of sentiment analysis?
We can use sentiment analysis to eg. analysis blog posts, chats social media texts or opinions about products.
3. Why and How Companies Should Use Sentiment Analysis?
Companies could use sentiment analysiss to check which posts are positive or negative about some products. Using collected information to improve offerts and mayby dismiss products that people dislike. They can check opinions about customer service to check customers feelings about offered service.
Hi Piotr, thank you for for your comment.
DeleteIt's great that you are familiar with the wide use of sentiment analysis in a business. Another application worth mentioned is identifying people facing depression based on their social media comments, or seeking for speech of hatred to blocked it efficiently.
1. Do you think is possible for an algorithm to analyze meaning of the text as accurately as human?
ReplyDeleteI think it all depends on the text. If the algorithm is to analyse simple messages, customer's feedbacks and social media posts full of emoticons, emojis, gifs, etc, then I would say "yes". But if we are talking about complex texts with metaphors, irony, overtones, hidden meanings and messages, then my is answer is "no". We're still not able to create a decent CAT tool to speed up the process of translation, so how can we create an algorithm analyzing meaning as accurately as human being?
2. What are the applications of sentiment analysis?
I guess sales and marketing or some dummy psychological tests on social media platforms. At the moment nothing else comes to my mind.
3. Why and How Companies Should Use Sentiment Analysis?
I agree with the previous speakers, using sentiment analysis by companies can have positive impact on sales if the company is striving to really satisfy its clients needs. One can also use it before launching new products onto to the market, to make trials on focus groups and receive good and reliable feedback when it comes to the item to be introduced and potential adjustments and improvements that it may require.
Thank you for your thought.
DeleteI'm not fully agreed with you that customers feedback is always a simple text. They can contain sarcasm or irony that make analyzing it a not trivial task
1. It seems to me that the continuous development of artificial intelligence can lead to such a situation. However, people often write sarcastically, put some hidden meanings, ambiguous expressions, which sometimes are hard to understand by human, let alone AI. If text is written without hidden messages, then most AI can analyze it in a way similar to a human being, looking for some patterns of expression.
ReplyDelete2. It can be used wherever a text appears, ie social media, comments and opinions of clients, responses in surveys, customer service, etc.
3. Thanks to the analysis of moods, the company can get an idea of how satisfied customers are with their products, they can get to know the real opinions about their new products. As a result, the company can make corrections to its products or to customer service if customers are dissatisfied with their work. Companies can use mood analysis also to promote new products that will go perfectly to customers' tastes, since we know their opinions on other products of a given brand.
Hi Monika thank you for your comments.
DeleteI think that e-commerce it's a great playground for sentiment analysis. Companies are investing a lot of money and a time of their data scientist to figure this out. The big break-through is waiting around the corner.
1. Do you think is possible for an algorithm to analyze meaning of the text as accurately as human?
ReplyDeleteFirst thing: "as accurately as human" doesn't necessarily mean it is accurately - humans tend to interpret facts in a way that does not collide with their world view.
Algorithms based on statistics have a natural limit that cannot be surpassed. To properly evaluate a text, a context is needed, and a whole corpus of knowledge of basic facts - the common knowledge.
2. What are the applications of sentiment analysis?
It can be used to get feedback on decisions made and processes used by companies, which allows to act towards desired outcome. Also, it might be a base of some early warning system, that would trigger an alert before some social media outcry happens. But it is all possible when subjects of analysis are not actively trying to counter it, like e.g. in China.
3. Why and How Companies Should Use Sentiment Analysis?
Companies should use Sentiment Analysis to monitor its perceived public image, and act accordingly. It can also be used by social media providers to weed out the most primitive cases of online harassment (but not more refined ones).
Hello Przemek, thank you for your thoughts.
DeleteVery interesting observation about early warning system for social media. I think that, that kind of alert will be valuable part of reputation management.
1. Do you think is possible for an algorithm to analyzed meaning of the text as accurately as human?
ReplyDeleteWhen I read this question first time my answer was "no way",but when I think about it more I have to answer "yes ... but not exactly". Technically similar to recursive neural networks, we analyze the text fragment by fragment. It's only later that we put it into full sentences and full stories.
2. What are the applications of sentiment analysis?
The most valuable knowledge could directly concern the ways in which listed companies take action. Forecasting trends on the market is knowledge all over the world.
3. Why and How Companies Should Use Sentiment Analysis?
In this way they can reject the contractor if the "mood" is uncertain and nervous which means uncertainty for the whole investment.
Hi thank you for your comments.
DeleteIt's interesting what you noticed about contractors, and fact that companies can used sentiment analysis not only in terms of customer relations, but also other type of business relations
1. Do you think is it possible for an algorithm to analyzed meaning of the text as accurately as human?
ReplyDeleteWell, it wouldn’t have the same accuracy as human analysis but it may be very similar. Much depends here on the ability to combine facts and the correct inference - deduction. Often the results of reasoning turn out to be paradoxical and with this AI will certainly have problems (since the man even has problems).
2. What are the applications of sentiment analysis?
The main application of sentiment analysis is recommender system. Moreover, it can be used in marketing or even in clinical medicine. I’m thinking about using sentiment analysis in music not only to classify songs but also to rate emotions.
3. Why and how companies should use sentiment analysis?
In my opinion business should use sentiment analysis because it gives fairly objective assessment of the business. I mean not only the analysis of customer ratings, but also a deeper analysis of customer interests or their emotional commitment to the brand. It is necessary to distinguish the opinions of customers from the emotions they convey. Here sentiment analysis can help a lot.
Hi Damian thank you for sharing.
DeleteFrom what I know recommendation systems do not use sentiment analysis per se they rather use association rules, you can look at basket analysis, In my opinion those algorithms are rely interesting.
It is used - in your link to Wikipedia it is explained in the point number 6 ("Application in recommender systems")
Delete1. Do you think is possible for an algorithm to analyzed meaning of the text as accurately as human?
ReplyDeletePersonally I am sure that it is not possible for the algorithm to analyze meaning of the text, especially relating to the specific product, as accurately as human. First of all, algorithms are based on the information given by the programmers. And we all know, that one person, or even few persons - if such algorithm is written by the group of people - can not predict all possible contexts or ways of expression. Moreover, sometimes it is all about emoticons or inside jokes and I think that it is not possible to be detected by today’s algorithms. Finally I am sure that nowadays AI and other virtual inventions can not guess the humans emotions as accurately as humans themselves.
2. What are the applications of sentiment analysis?
As it seems to me, the main application of sentiment analysis is to check how the company’s products are accepted by the consumers. By using sentiment analysis, traders can make improvements to their products, or even withdraw them from purchase. It can also be used in case of deciding if some TV series or reality shows are popular and if they deserve to be continued or if they should be cancelled. Also politicians can apply sentiment analysis during their campaigns to see what they should do to ensure the victory in the election. Even scientists may take advantage from the sentiment analysis, when it comes to the way of describing their findings - of course the results will remain the same, but sometimes for the usual people it is all about the method of presenting some studies and such methods can be deciding factor if the findings will gain common acceptance or not. But this are just some ideas that right now came to my mind, I am sure that there are a lot more fields, where the sentiment analysis can be applied.
3. Why and How Companies Should Use Sentiment Analysis?
As I mentioned above, companies should use sentiment analysis to check if their products are popular, what are advantages and disadvantages of them and what can be change to provide the maximum level of consumer’s satisfaction, hence the highest sell. As it concerns the way how the companies should use described method, I am no expert on that field ;) of course they should focus on the opinions presented on social media or on the website of producers. They should also pinpoint the group of consumers who are the target of selling specific product and search on their social media any opinions or even videos about the products. Also before selling an item the company can check what are the moods about it in two ways - first of all, if any other company is selling similar product, they can focus on opinions about such product, and secondly - they can post on their website that they are planning to bring in the new product and check what consumers have to say about it.
Hello Łukasz thank you for your comprehensive comment.
ReplyDeleteI only want to emend that during writing of sentiment analysis algorithms programmers use already build library prepared for most popular languages.
Klara,
ReplyDeleteI’ve got familiar with your links with a real interest.
1. Definitely I do not believe in algorithms’ analysis of similar accuracy than humans’. Sentiment analysis is the automated process of understanding an opinion about a given subject from written or spoken language. “Automated” means that it always follows human, of course it can be the result of self-teaching procedures hided in the special computer software, but even with assumption of its sophisticated structure and predictions, it will always be only artificial analysis. The opposite cannot be proven ;-)
2. Currently, sentiment analysis is a topic of great interest and development since it has many practical applications. Since publicly and privately available information over Internet is constantly growing, a large number of texts expressing opinions are available in review sites, forums, blogs, and social media. There are many types and flavors of sentiment analysis and SA tools range from systems that focus on polarity (positive, negative, neutral) to systems that detect feelings and emotions (angry, happy, sad, etc) or identify intentions (e.g. interested v. not interested).
3. With the help of sentiment analysis systems, the unstructured information gained by companies, could be automatically transformed into structured data of public opinions about products, services, brands, politics, or any topic that people can express opinions about. This data can be very useful for commercial applications like marketing analysis, public relations, product reviews, net promoter scoring, product feedback, and customer service.
Many thanks for interesting issue,
BR,
Marta
Thank you so much for this nice information. Hope so many people will get aware of this and useful as well. And please keep update like this.
ReplyDeleteText Analytics Software
Text Analytics NLP