【分手或和好,让电脑程序来告诉你吧】

【分手或和好,让电脑程序来告诉你吧】南加州大学和犹他大学的研究小组设计了一款电脑程序,能够预测两口子的关系变差或者更好。它以夫妻之间相互交流时的语调作为研究对象,记录了100对夫妻的交流语音并对双方的地位进行了长达5年的追踪调查,其结果准确率达到79%,甚至比专家更为有效。

brad pitt jennifer aniston

Researchers have created a computer program that can predict relationship success with better accuracy than therapists.

The program is capable of predicting whether you and your spouse will have an improved or worsened relationship over time with 79% accuracy. It even fared better than relationship experts, who kept comments on married couples' therapy sessions.

The program works by analyzing the tone of voice used when couples spoke to each other. It was designed by a team hailing from University of Southern California's school of engineering and the University of Utah.

“What you say is not the only thing that matters, it’s very important how you say it," Md Nasir, a doctoral student at USC and author on the study, wrote in a press release.

To create the program, researchers recorded hundreds of conversations from 100 couples in marital therapy sessions over a two-year span. They then tracked each couples' martial status for five years.

The researchers broke each of the recordings down into acoustic features and looked at factors like pitch and voice intensity. It was important to analyze multiple conversations between couples so the researchers could fine-tune the program to see how the couples' vocal acoustic features changed over time, and what that said about the relationship.

Eventually, the researchers hope to analyze body language and the actual words spoken to predict relationship outcomes.

“Looking at multiple points in time and looking at both the individuals and the dynamics of the (duo) can help identify trajectories of the their relationship," Panayiotis Georgiou, an assistant professor of electrical engineering who co-led the study, wrote in the press release.

http://www.techinsider.io/usc-university-of-utah-created-a-computer-program-that-can-predict-relationship-success-2015-11


Comments are closed.



无觅相关文章插件