Improving Natural Language Recognition for Leading Social Media Firm
High-Quality Data Collection Improves Machine Learning Algorithm
A leading social media company needed to better understand user-generated content from natural language. It required large amounts of data to improve its machine learning model. Training the model with false positives and negatives was also a requirement. Data needed to be sourced on a tight timeline to hit internal goals and an external release date.
The social media company reached out to Appen for assistance on this collection project. Appen recruited hundreds of participants in just a few days – and in 2 months, more than one million samples were collected.
Not only did the data collected improve the platform’s help functions, ads, videos, and more, the firm met its deadlines and budget requirements. With Appen’s diverse geographic and demographic rater pool, there were ample language variations for data scientists to rely on one data set for the whole process.