12.8 C
Nicosia
Thursday, November 30, 2023

Recommendation systems and economy of attention

Must read

& Sigma; & upsilon; & sigma; & tau; or & mu; & alpha; & sigu & tau; & alpha; & upsilon; & sigma; & tau; ά & sigma; & epsilon; & omega; & nu; & kappa; & alpha; & iota; & omicron; & iota; & kappa; & omicron; & nu; & omicron; & mu; ί & alpha; & tau; & eta; & sigmaf; & pi; & omicron; & chi; or & sigmaf;

POWERED BY:

 & Sigma; & upsilon; ; & tau; or & mu; & alpha; & tau; & alpha; & sigma; & upsilon; & sigma; & tau; ά & sigma; & epsilon; & omega; & nu; & kappa; & alpha; & iota; & omicron; & iota; & kappa; & omicron; & nu; & omicron; & mu; ; & eta; & sigmaf; & pi; & rho; & omicron; & sigma; & omicron; & chi; or & sigmaf;

06:26 Saturday
March 12, 2022 INSIDER • GOAL • IN-CYPRUS • PROPERTY • CYPRUS-IS • CAREER LIKE • GOING OUT • XRYSES SYNTAGES • ACTIVE • PHILGROUP NEWS & nbsp; SOCIETY & nbsp; INSIDER & nbsp; OPINIONS & nbsp; GOING OUT & nbsp; CULTURE & nbsp; SPORTS & nbsp; AUTO & nbsp; GOOD LIFE & nbsp; USEFUL & nbsp; PRINTED EDITION & nbsp;

& Sigma; & upsilon; & sigma; & tau; or & mu; & alpha; & tau; & alpha; & sigma; & sigil; & upsilon; ά & sigma; & epsilon; & omega; & nu; & kappa; & alpha; & iota; & omicron; & iota; & kappa; & omicron; & nu; & omicron; & mu; ί & alpha; & tau; & eta; & sigmaf; & pi; & rho; & omicron; & sigma; & sigma;

& Sigma; & upsilon; & sigma; & tau; or & mu; & alpha; & tau; & alpha; & ta; & sigma; & sigma; ά & sigma; & epsilon; & omega; & nu; & kappa; & alpha; & iota; & omicron; & iota; & kappa; & omicron; & nu; & omicron; & mu; ί & alpha; & tau; & eta; & sigmaf; & pi; & rho; & omicron; & sigma; & sigma;

 & Sigma; & upsilon; & sigma; & tau; or & mu; & alpha; & tau; & alpha; & sigma; & upsilon; & sigma; & tau; ά & sigma; & epsilon; & omega; & nu; & kappa; & alpha; & iota; & omicron; & iota; & kappa; & omicron; & nu; & omicron; & mu; ί & alpha; & tau; & eta; & sigmaf; & omicron; & sigma; & omicron; & chi; or & sigmaf;

ARCHIKIEIDISEISPOLITIKITOPIKAPARASKINIOMEDIAELLADAKOSMOSEPISTIMIPERIVALLONTECHNOLOGIASYNENTEFXEISCHRYSES ETAIREIESKOINONIAEIDISEISKYPROS OF CHTHESANAGNOSTI HAVE LOGOEPISTOLESSYNENTEFXEISDIMOFILI IN INTERNETVIDEOSAFIEROMATAADVERTORIALINSIDEREPICHEIRISEISKYPROSBRAND VOICECAPITALFORBESBLOOMBERG OPINIONPROTAGONISTESKAFTI GRAMMICHRYSES ETAIREIESKARIERAAPOPSEISARTHRA IN “F” INTERVENTION IN “F” MESSAGE YOUR SKITSAO WORLD TWITTERGOING OUTTHINGS TO DOCINEMAMOUSIKIESTIATORIABAR/CAFETV ODIGOSTILEORASIPOLITISMOSKYPROSKOSMOSKRITIKESEKDILOSEISPROSOPAATHLITIKAPODOSFAIROMPASKETPARASKINIAELLADADIETHNIALLA SPORAPOPSEISSKITSOVIDEOSAUTONEAPAROUSIASIAPOSTOLESAGONESKALI ZOIFGEIADIATROFIEF ZINASTRACHRYSES SYNTAGESSYNTAGES SEFVIMA VIMACHRISIMAFARMAKEIAGIATROIAERODROMIALIMANIATILEFONAOPAPKAIROSSYNALLAGMALACHEIAAPPSPROSFORESENTYPI EDITION POLITIKITOPIKAPARASKINIOMEDIAELLADAKOSMOSEPISTIMIPERIVALLONTECHNOLOGIASYNENTEFXEISCHRYSES COMPANIES EIDISEISKYPROS OF CHTHESANAGNOSTI HAVE LOGOEPISTOLESSYNENTEFXEISDIMOFILI THE INTERN ETVIDEOSAFIEROMATAADVERTORIAL EPICHEIRISEISKYPROSBRAND VOICECAPITALFORBESBLOOMBERG OPINIONPROTAGONISTESKAFTI GRAMMICHRYSES ETAIREIESKARIERA ARTICLES IN “F” INTERVENTION IN “F” MESSAGE YOUR SKITSAO WORLD OF TWITTER THINGS TO DOCINEMAMOUSIKIESTIATORIABAR/CAFETV ODIGOSTILEORASI KYPROSKOSMOSKRITIKESEKDILOSEISPROSOPA PODOSFAIROMPASKETPARASKINIAELLADADIETHNIALLA SPORAPOPSEISSKITSOVIDEOS NEAPAROUSIASIAPOSTOLESAGONES YGEIADIATROFIEF ZINASTRACHRYSES SYNTAGESSYNTAGES SEFVIMA STEP FARMAKEIAGIATROIAERODROMIALIMANIATILEFONAOPAPKAIROSSYNALLAGMALACHEIAAPPSPROSFORES ENTERPRISES LTD BRAND VOICE CAPITAL FORBES BLOOMBERG OPINION PROTAGONISTS HOT LINE GOLDEN CAREER COMPANIES

& Sigma; & upsilon; & sigma; & tau &; & sigma; & upsilon; & sigma; & tau; ά & sigma; & epsilon; & omega; & nu; & kappa; & alpha; & iota; & omicron; & iota; & kappa; & omicron; & nu; & omicron; & mu; ί & alpha; & tau; & eta; & sigmaf; & pi; & rho; & omicron; & sigma; & omicron; & chi; or & sigmaf;

EXCLUSIVE COOPERATION

& Sigma; & upsilon; & sigma; & tau & alpha; ; & sigma; & upsilon; & sigma; & tau; ά & sigma; & epsilon; & omega; & nu; & kappa; & alpha; & iota; & omicron; & iota; & kappa; & omicron; & nu; & omicron; & mu; ί & alpha; & tau; & eta; & sigmaf & r ; & sigma; & omicron; & chi; or & sigmaf;

TOGETHER

& Sigma; & upsilon; & sigma; & tau; or & mu; & alpha; & tau; & alpha; & ta; sigu; & sigma; ά & sigma; & epsilon; & omega; & nu; & kappa; & alpha; & iota; & omicron; & iota; & kappa; & omicron; & nu; & omicron; & mu; ί & alpha; & tau; & eta; & sigmaf; & pi; & rho; & omicron; & sigma; & sigma;

Powered by

& Sigma; & upsilon; & sigma; & tau; or & mu; & alpha; & tau; & alpha; & ta; sigu & sig; & upsilon; ά & sigma; & epsilon; & omega; & nu; & kappa; & alpha; & iota; & omicron; & iota; & kappa; & omicron; & nu; & omicron; & mu; ί & alpha; & tau; & eta; & sigmaf; & pi; & rho; & omicron; & sigma; & sigma;

 & Sigma; & upsilon; & sigma; & tau; or & mu; & alpha; & tau; & alpha; & sigma; & upsilon; & sigma; & tau; ά & sigma; & epsilon; & omega; & nu; & kappa; & alpha; & iota; & omicron; & iota; & kappa; & omicron; & nu; & omicron; & mu; ί & alpha; & tau; & eta; & sigmaf; & omicron; & sigma; & omicron; & chi; or & sigmaf;

Next Recommendation Systems and Attention Economy HOME • INSIDER • CYPRUS • Recommendation Systems and Attention Economy

 & Sigma; & upsilon; & sigma; & tau; or & mu; & alpha; & tau; & alpha; & sigma; & upsilon; & sigma; & tau; ά & sigma; & epsilon; & omega; & nu; & kappa; & alpha; & iota; & omicron; & iota; ; & nu; & omicron; & mu; ί & alpha; & tau; & eta; & sigmaf; & pi; & rho; & omicron; & sigma; & omicron; & chi; ή & sigmaf;

The age we are going through, incomprehensibly, is characterized as the age of information. At any given time, huge amounts of information are generated and circulated on the internet. Indicatively, it is worth mentioning that every minute that passes, 60 hours of new material is added to the youtube platform . At the same time, huge amounts of information are circulated on social media, information websites, platforms for the production and distribution of multimedia material, as well as in online stores. And all these means try to simultaneously claim the limited duration of consumer attention. This competition produces a new ecosystem, known as the “attention economy”. The resources of this economy are the users, and the electronic services compete in order to gain their attention and time. The more time we spend on their products, the more power they gain.

Write

Dr. Melina Moleski, collaborating teaching staff, department of informatics, University of Naples Paphos

Dr. Savvas Hadjichristofis, Associate Professor of Artificial Intelligence, University of Naples Paphos, member of the SCIENCE HOAXES team

Many times, for these companies, the users are the product itself. For example, the more visitors, and & nbsp; The more time they spend on Facebook, the more enticing the platform becomes for advertisers. As a result, companies are investing millions in the production of a digital dopamine system, which leads to the addiction of users to their services. And one of these digital dependency generators is Recommender Systems.

WHAT ARE RECOMMENDATION SYSTEMS AND HOW DO THEY WORK?

Referral systems are actually a special class of algorithms used to suggest choices to online service users. It is in the system that tries to personalize the available services of a platform in order to improve the feeling of user satisfaction. For each of us, the content we see on the main page of Youtube is different, and according to our preferences. The products offered by Amazon, the ads we see on Facebook, the movies selected by Netflix, the songs that make up our Spotify list, all carefully selected to satisfy our choices. The smart algorithms try to map in detail the profile of our preferences and provide us with the most suitable options, keeping us trapped in the web of the companies they serve. The better their suggestions, the more time we devote to the service. And as time goes on, so does our addiction.

The technology behind the principle of operation of these algorithms varies. Their efficiency can not be easily evaluated. If one tries to divide the recommendation algorithms into general categories, one will easily find that 2 styles can be formulated. The first style adopts the principles of “collaboration-based filtering” while the second adopts the principles of “content-based filtering”.

The principle of operation of “collaboration-based filtering” algorithms is based on the assumption that people with similar choices may have common interests. Each service that uses this form of algorithm maintains a user profile, in which it records its selections, ratings, and navigation history on the service. Algorithms try to identify users with similar options and sort them into groups. In this way, the choices of one member of the group, become suggestions for the other members. Suppose, for example, a group of Amazon Prime users who watched the same series and movies, leaving similar ratings. These users, based on the algorithm, form a group. If a new entry on the platform attracts the interest of a large portion of the team members, then it automatically becomes a proposal for the other members. Given the large number of subscribers, the platform can easily map people with a similar profile and form groups with relatively high similarity in the selection criteria.

In contrast, “content-based” filtering algorithms base their proposals on the content itself. For each user of the platform, the algorithm configures the profile of his preferences. Then, for each entry that enters the platform, the algorithm analyzes its content and presents it as a suggestion only to users who may be interested. Suppose for example the platform of a service that offers online music. The platform knows for each of its subscribers, the types of music that interest him (rock, pop, etc.). This information can be entered by the user himself or can be obtained by following his current options. For each new track that goes up on the platform, the algorithm analyzes either the data that accompanies it (artist, record company, etc.) or the track itself (evaluating the rhythm or style of the song) and sorts it into a specific kind of music. The result of this evaluation feeds the platform with information, able to identify & nbsp; to which subscribers it can recommend it.

It is worth noting that the rapid development of artificial intelligence in recent years, has led to the development of highly intelligent and efficient recommendation models that use techniques from both families of algorithms. An example of a good hybrid algorithm is used by the Netflix platform. The suggestions to the subscribers of this service & nbsp; are based both on the choices of the users with a similar profile and on the personal characteristics of the user, as they were formed by his evaluations so far.

WHY THEY ARE USEFUL FOR THE CONSUMER?

Some of us will remember that Netflix (when it was founded in 1997) was originally a mail-order DVD rental company, until it switched to streaming service in 2007. What few of us know is that, to attract customers to subscribe to the new service, Netflix people ran the following experiment: Customers were asked “What would you like to know more about before you sign up for Netflix?”. The most popular answer (46% of the answers) was “knowledge of all available movies and TV shows”. ‘So the platform decided to show customers all the available content on the homepage. But the results were not what they expected. The appearance of too large a volume of content confused the viewers making the choice of film even more difficult. Many of them browsed without ever registering.

Netflix finally redesigned the experiment. This time the platform was designed to offer a sneak peek – but not an overview – of the content. They used an image that hints at an extensive list by topic, not allowing visitors to browse the long list. This design increased the interest of the visitors, at the same time enhancing the possibility of converting them to subscribers in the streaming service. This is how Netflix tackled the phenomenon of “choice overload.”

This phenomenon, also known as “choice paralysis” or “choice paradox”, describes how people find it difficult to choose as the number of choices grows. This is due to limited cognitive resources. Therefore, the more options we have to consider, the more our cognitive systems are burdened and our energy (decision fatigue) is drained. The “overload of choices” often paralyzes us, prevents us from making any decision at that moment (action paralysis). At the same time, research shows that a large volume of options is associated with lower confidence in our choice, with a greater likelihood of regretting our decision and with reduced satisfaction in general. >

Recommendation systems transform human choice and free us from “overload of choices”. A study by McKinsey shows that 35% of the products consumers buy on Amazon and 75% of those who watch Netflix rely on recommendations from such systems. At a time when “best choice” is less important than “pretty good choices”, recommendation models have been incorporated into our daily lives for good. They increasingly influence our choices in clothing, entertainment, food and medicine. They also affect the texts we send, the friends we communicate with, the potential clients we prioritize, the experts we seek, the candidates we hire, the investments we choose and the schedules we follow.

LinkedIn collects data from users' profiles about their education, specialty, experience and skills. At the same time, it suggests to the user jobs that he may be interested in, as well as people whose utilization can bring added value to companies. In another example, Gmail suggests entire business email phrases based on previous chat text in older emails. Each week, Spotify creates a new personalized playlist of 30 songs for each subscriber based on their latest music choices. Amazon and other online stores offer additional products based on the choices of other customers who purchased the same product. Tinder algorithms detect a user's swipe left, swipe right patterns and use them to predict possible physical attraction between couples, thus shaping future suggestions to users. Humu provides consulting services for the use of referral systems to improve the efficiency of human resources. IBM and Salesforce use and offer such systems as in-house productivity platforms for campaign design and targeted sales.

CONCLUSIONS

Undoubtedly, the survival of online services depends on the efficiency of the referral systems they use. The future of these systems, however, lies in challenges, perhaps greater in the formulation of universal propositions, even in the absence of an immediate search. Imagine an afternoon walk in the mall. Suddenly, your mobile phone informs you that in the store that is in your right hand, there is an offer for the mobile phone that you just searched for a few days ago, expressing interest, in various websites.

It is important to emphasize the use of personal data. The value that comes from reliable recommendations (in terms of saving time, energy and gray matter, but also satisfaction from the final choice) justifies in the minds of (most) consumers the use of personal information – as long as transparency and confidentiality are maintained.

From February Insider Magazine

Source: www.philenews.com

- Advertisement -AliExpress WW

More articles

- Advertisement -AliExpress WW

Latest article