In a 90 sec video here I have provided the general high-level introduction of machine learning.
In this article, I will explain unsupervised learning and how it is used to solve everyday problems by business owners.
Unsupervised Learning is the type of machine learning where the learning process is not supervised. In this case, we provide the data to the model without any specific labels or outputs. The model will then find patterns in the data itself by finding similar attributes between different data elements. When the data is arranged in groups based on similar attributes or properties, it is called Clustering. Anomaly detection is another type of unsupervised learning where we detect anomalies within the data based on the properties of the data. Dimensionality reduction is another topic that comes under unsupervised learning, but we will not cover it in this post.
Depending on the attributes that are of importance different elements of the data can be arranged into different clusters. For example, in the picture below the fruits and the vegetables are mixed. After feeding our model with this mix data of fruits and vegetables, in this case, model choses to take color as the main feature and creates clusters accordingly.
Since unsupervised learning is all about finding the patterns within the unlabeled data, one application or type of it is also anomaly detection. When the algorithm is given constantly similar data, it will be easy to identify pattern or data input which does not share the similar behavior or attributes with rest of the data. In the time series data this is usually used in predictive maintenance use case. Similarly, anomaly detections is also used to identify uncharacteristic behaviors from a video feed in security use case.
Recommender systems are widely deployed across various industries. Clustering is essentially the major type of unsupervised learning that is used in the recommender systems. Either it is your e-commerce store (eBay or Amazon), your entertainment platform (YouTube or Netflix) or your social network (Facebook or LinkedIn), they all use recommender systems to find the patterns and similarities to recommend you either products to buy, videos/movies/series to watch or friends/connections to add. As shown in the example below, the recommender engine from Netflix is recommending a user “series to watch” because they share similar attributes of the series, he/she has just finished watching.
To learn more about the solution using machine learning to solve real world use-cases visit Tech Data’s IoT solutions catalogue page here.
IoT Solution Architect (A.I and Vision)
IoT and Analytics Team
Tech Data Europe