Machine learning for personal finance management
Kosta, one of our Developers at Numbrs, explains the fundamentals of machine learning and how we use this technology to provide a more custom-tailored approach to personal finance…
Numbrs is much more than just a bank aggregation app. Getting an instantaneous and clear overview of one’s accounts has, of course, significantly increased our users’ awareness and understanding of their finances. But is this enough?
When looking at a person’s bank statements, we find that certain transactions are repeated periodically in a very predictable fashion, like one’s salary or rent, whereas others have a more erratic pattern, such as food expenses, car repairs or vacations.
The overall picture, however, is a coherent one, which conforms to an underlying pattern. It’s a bit like listening to music: There is a pulse, a rhythm as well as some space for a solo and improvisation. Detecting periodicity in transactions, in other words, is not that different from detecting the beat of a song. Each type of transaction has its own timbre, like a musical instrument, all playing together to form each user’s individual financial song.
Becoming a mobile personal financial advising app requires listening to each user’s intimate song and letting them know when they start getting off-rhythm. Once the rhythm is known, we can start making predictions about a user’s future transactions and thus help them to start planning ahead.
This is why our growing Data Science Team is at the core of our operations. From the automatic categorisation of transactions to future expense predictions, we strive toward non-supervised self-learning processes that provide a new vision for financial account management.
Please click here to read our impress: Link