Use of AI to Predict Relationship Satisfaction

Can machines love? Perhaps this question will remain unanswered till unknown future. But one thing is certain, Artificial Intelligence based machine can calculate and predict if you are in love. This idea is so amazing that how can a machine that has no emotions, no free will can determine the love between two individuals. Thank to artificial intelligence that uses certain pieces of information to determine whether love is real or fake.

Read more on How AI supports love of life.

It’s not as easy as it looks to understand a successful relationship. It depends on several factors which differ from individual to individual. People make relationships who are two different souls, and when they find comfort-ability, they join with each other. What makes them happy together is the same thinking approach and loving the same things to do in their lives ahead. People who have a comfortable and long-lasting relationship are those who care for each other, not by words but in reality. They stand by to support their partner in every situation, either its time to take important decisions or to recover any illness. It’s all about the real relationship which lasts till the last breath of a partner.

Read more on How Japan is using AI to help staff find love.

But in contrast, there are some relationships which look real but no more than a fake story leading to divorces. These external relations are made when both partners think to live together but didn't notice their habits. Consequently, they start observing differences and fighting to prove their point of view to be right. It leads to a relationship ending; in the United States, 40% of marriages end with a divorce even after a year.

There might be two factors which result in people’s engagement:

1. How People Approach Their Potential Lover?
The platform used for introduction and communication is significant in determining the life of a relationship. In the present era, people connect with others through mobile phones, tablets, laptops, and computers. Physical meetups come with the desire of a member after knowing each other well on different online platforms. Emojis help in sharing real-time mood swings among lovers that boost their feelings.

2. Who Is the Right Person as A Partner?
People come across a variety of humans, and not everyone attracts them. Some characteristics in others make persons feel about them. It's a terrifying condition where no one proposes other, but machines help in approaching that particular person. AI algorithms help in finding the perfect match with whom one can feel compatible.

The question is how to identify are you in a relationship with the right person or not? Artificial intelligence is helping people to choose their partners by placing their thoughts and intends. Several companies are struggling hard in developing software and apps to distinguish the right relationships from the fake ones.

Recommendation Engines

Recommendation engines are working through several apps for helping the users to see content according to their first search. Amazon also offers this system to suggest its customers with the related products. Moreover, Netflix suggests similar videos to a user's interest to engage it for significant periods. Their functionality can be described as:

a) Historical Trends

It represents a person's past interests to predict his future liking and disliking. A person's past searches help the algorithm to give suggestions in the recommendation bar.

b) People Searching Same Data

Algorithms combine people in groups based on their similar searches and assign them a specific category. It defines similarity among people on their buying and browsing behaviour, their age and genders. It differs from historical trend method as it shows the result of other persons as well of the same group.

c) Content-Based

It doesn't focus on a person's intend but on particular products. For instance, a movie providing website recommends films based on genre, language, length, and cast of the movie.

Some recommendation engines work as a hybrid and save all kinds of queries to show for the next time in suggestions.

After collecting the data, AI algorithms become active to share likelihood information about a particular person. It's based on their interest; the algorithm also updates its suggestions when a user accepts or rejects a specific request.

1. Artificial Intelligence Guessing A Relationship Status

DataRobot an AI organization, launched a tool that uses data from Stanford University and asks six questions related to one's relationship. It uses this data to predict the next years of a relationship; will it go further or end?

How DataRobot Labs Arm Develop this Tool?

Stanford University made a quiz in 2009 about 4000 Americans known as ‘’How Couples Meet and Stay Together.'' It collected information from many couples, and then after some years, repeated the same quiz and shared the collected data with the public. The primary function of this predictive model was to predict outcomes. The researcher used this data to develop a model for predicting relationship statuses.

The director of DataRobot Labs, Greg Michaelson, shared his views in an interview:

‘’That’s kind of got the icky feel to it. we wanted to stay away from anything too personal and sensitive.’’

The company asks only six questions that are not too personal, and couples don't hesitate to respond to them. Even the app doesn't ask if partners are living together or not.

Six Questions It Asks

• What is your relationship status?
• What’s the highest level of education completed?
• How old are you and your partner?
• How long have you and your partner been together?
• How many the children between the ages of 2 and 5 live with you?
• On average, how many different relatives do you see each month?

2. Western University Research Project

According to a study, relationship satisfaction is necessary for carrying a better and long-term relationship. For this sake, Western University used a technical approach to find the answer. They conducted an Artificial Intelligence experiment to study the status of Romantic Relationships. They took data from 11,196 couples, and they were from 43 self-reported datasets. Then this data was inserted to Machine Learning algorithm.

After feeding data, the algorithm scanned it to discover particular patterns supporting successful relationship. AI declared that the right prediction about a happy romantic relationship is the commitment of partners' spouse. Moreover, it also suggested that for a happy relation, appreciation, closeness, fight resolution, and sexual comfortability is necessary.

The psychology professor of Western University Samantha Joel led this study, stated that:

“In other words, relationship satisfaction is not well-explained by your partner's self-reported characteristics.”

He added further:

“Satisfaction with romantic relationships has important implications for health, well-being, and work productivity. But research on predictors of relationship quality is often limited in scope and scale, and carried out separately in individual laboratories.”

3. University of California’s Research Project

Paul Eastwick, Joel, conducted this machine-learning study in the University of California, Davis including 84 other scholars. They took data from more than 11,000 couples and 43 self-reported databases about romantic relationships. Proceedings of the National Academy of Sciences published this study to respond to a persistent question: “what predicts how happy I’ll be with my relationship partner?”

4. Social Media App, Facebook Helping in Predicting Love & Breakup Status

Facebook is one of the online platforms which collects data from its users in predicting either they're in a relationship or not. OKTrends was famous for helping people in optimizing the message they receive from opposite or similar gender. Then Match.com bought original website of OkCupid and OKTrends stopped posting additional content.

Facebook comes with more reliable and real data than dating sites. It joins millions of people together who make relationships and provide essential representative data to predict valuable results. They found some interesting facts, as given below.

Take an interest in Your Partner's Interests.

Lars Backstrom from Facebook and Jon Kleinberg from Cornell University concluded that the best lovers engage with each other. However, both don't need to have familiar friends; vital thing to match is their interests in the same activities.

• Couples having familiar friends range up to 24.7 per cent.
• More common relationship (up to 50 per cent) occurs in social circle friends (sports club, drinking friends, colleagues, co-workers)
• Couples having more common pictures together and checking the activity of their partner exceeds results up to 70 per cent.

It helps in predicting breakups couples stop appearing in pictures together for long periods. Moreover, they also stop liking and commenting on each other's posts. They obtained data from not only a single couple but the study includes all couples on Facebook and their behavior.

Interestingly, they found that heterosexual couples are mostly of the same age and get older together. Homo-sexual couples show age gaps of about 4.5 years in their interactions after 38 years. In western countries, people love to marry a partner with the same age. The life of a particular relationship depends on how long a couple stays together. According to a Facebook Study Report:

“About half of all Facebook relationships that have survived three months are likely to survive to four years or longer.”

The book, The Facebook Effects mentioned algorithm developed by Zuckerberg at Harvard to predict which of his friends stay connected. The algorithm worked with 33 per cent accuracy. He created it for fun with his friends about their social life and interactions.

5. StayGo App and its Accuracy

A team of psychology and relationship experts developed StayGo app in 2016. It asks 20 questions from the couples for determining their compatibility, comfortability, and money handling. Then it assembles result sin "SG Score" out of 100. But it's still unknown how accurately its algorithm works.

Ending Words

It’s a challenging task to predict relationship status by technology and artificial intelligence. Furthermore, it works on some general factors that might fail to indicate all human's real emotions about a particular person. Some variable factors, including environmental and biological attraction, have a noticeable impact on a relationship. Scientific Advisor at dating site Match, Justin Garcia said:

“Relationships are dynamic processes and complex system.”

AI will change the way completely in predicting relationship that will help in making the right connections. AI is helping predicting sleeping, exercising, and financial pattern of a particular person. Soon, AI in Healthcare introduces a way to stop the state of a broken heart. It will also help in reducing loneliness by providing the right match to people; thus. Reducing stress and depression in the young generation.