MeeMix Taste Prediction Platform
MeeMix combines innovative thinking and vast professional experience into ground-breaking, commercially-applicable taste prediction technology
MeeMix hybrid technology integrates a number of independent methodologies in order to predict musical taste after as little as 2 actions taken by a user and at optimal accuracy. The combination of multiple methods, such as signal processing, machine learning, behavioral analysis and collaborative filtering, marks an innovative, unique approach, which technologically distinguishes MeeMix from other solutions, and provides a significant commercial edge.
The MeeMix Taste Prediction Platform serving as the technological foundation for the various MeeMix services, is powered by 3 main components:
- Off-line Setup - Automatic Classification
Our generic Auto Classification engine automatically runs through our partner's musical catalog, classifying each musical piece by analyzing its musical attributes. Auto Classification can handle music in any language or style, allowing us to serve businesses across the globe. Auto Classification makes periodic updates to your catalog a quick and cost-effective process. Further, we currently offer the only technology capable of automatically classifying music mood.
The Auto Classification engine relies on an automated algorithm developed by statisticians and musicologists on our research team. The algorithm extracts hundreds of physical parameters from each audio file (song), and combines statistic and probabilistic modeling with empirical metadata analysis in order to assign valuable musical attributes to each such song. In essence, we cost-effectively break down each audio file into its physical components, and use these to accurately predict its musical genre, mood, BPM, and various other features. For example, Led Zeppelin's Stairway to Heaven will be assigned hundreds of parameter values reflecting its BPM, mood, pitch and other qualities.
It should be noted that as opposed to every other technology in the market, we offer the only methodology that classifies mood automatically and accurately, using our unique Mood Control module.
Since the Auto Classification engine relies mostly on statistical modeling, it features generic applicability that allows us to classify each of our partners' own musical catalogs, regardless of their size, musical genres or language.
- Real Time - Taste Prediction
Our unique hybrid technology automatically analyzes, acquires and continuously adjusts the musical taste of each customer in our partner's database, in real-time. It takes into account a variety of sources, such as customer mood, behavior and available information, collective taste and collaborative filtering.
The MeeMix Taste Prediction Platform was developed to automatically analyze and acquire the musical taste and mood of each user in our partner's database.
We found that musical preference is composed of elements of both nature and nurture. As such, we first use the song/artist initially requested by the user, as well as available information about the user (such as age and geographic location), in order to pinpoint their primary position in the musical hyperspace. Using an algorithm developed by statisticians specializing in psychology and human behavior, we then monitor the user's ongoing feedback to songs they experience, deriving more information about their current mood and which musical attributes they like or dislike, and adjusting accordingly our perception of their taste. While other technologies on the market incorporate some of these methods to some extent, MeeMix's is the only one to actually converge all these sources. Additionally, The MeeMix Taste Prediction Platform is the only available technology that automatically picks up on the customer's current mood without the need for specific, customer-initiated definitions.
In short, we first use information about who the user is and what they like, and then we continually fine-tune our insight into their musical taste. For example, we may analyze a 17-year-old English male teenager's positive ratings of various rock songs and derive conclusions about his mood at the moment and specific musical likes and dislikes.
-
API Interface - Simple, Cost-Effective Integration
A comprehensive API web service provides our partners with a seamless, cost-effective and automatically-updated integration to their applications, with no need for on-site implementation.
Thus, the multi-faceted MeeMix Taste Prediction Platform enables our partners, whether in cable, internet or mobile, to effectively combine our abilities to map their content catalogs and pinpoint their users' preferences into a personalized user experience. This, in turn, provides our partners with a critical advantage over competition: it significantly enhances customer experience and loyalty, while at the same time optimizes the conditions for customer acquisition and consumption and maximizes utilization of the provider's content.
