Solving Real-World Problems with Nearest Neighbor Algorithms

By Lillian Pierson

Hierarchical clustering algorithms — and nearest neighbor methods, in particular — are used extensively to understand and create value from patterns in retail business data. In the following paragraphs are two powerful cases in which these simple algorithms are being used to simplify management and security in daily retail operations.

Seeing k-nearest neighbor algorithms in action

K-nearest neighbor techniques for pattern recognition are often used for theft prevention in the modern retail business. Of course, you’re accustomed to seeing CCTV cameras around almost every store you visit, but most people have no idea how the data gathered from these devices is being used.

You might imagine that there is someone in the back room monitoring these cameras for suspicious activity, and perhaps that is how things were done in the past. But today, a modern surveillance system is intelligent enough to analyze and interpret video data on its own, without a need for human assistance.

The modern systems are now able to use k-nearest neighbor for visual pattern recognition to scan and detect hidden packages in the bottom bin of a shopping cart at check-out. If an object is detected that’s an exact match for an object listed in the database, then the price of the spotted product could even automatically be added to the customer’s bill. While this automated billing practice is not used extensively at this time, the technology has been developed and is available for use.

K-nearest neighbor is also used in retail to detect patterns in credit card usage. Many new transaction-scrutinizing software applications use kNN algorithms to analyze register data and spot unusual patterns that indicate suspicious activity.

For example, if register data indicates that a lot of customer information is being entered manually rather than through automated scanning and swiping, this could indicate that the employee who’s using that register is in fact stealing customer’s personal information. Or if register data indicates that a particular good is being returned or exchanged multiple times, this could indicate that employees are misusing the return policy or trying to make money from doing fake returns.

Seeing average nearest neighbor algorithms in action

Average nearest neighbor algorithm classification and point pattern detection can be used in grocery retail to identify key patterns in customer purchasing behavior, and subsequently increase sales and customer satisfaction by anticipating customer behavior. Consider the following story:

As with other grocery stores, buyer behavior at (the fictional) Waldorf Food Co-op tends to follow very fixed patterns. Managers have even commented on the odd fact that members of a particular age group tend to visit the store during the same particular time window, and they even tend to buy the same types of products.

One day, Manager Mike got extremely proactive and decided to hire a data scientist to analyze his customer data and provide exact details about these odd trends he’d been noticing. When Data Scientist Dan got in there, he quickly uncovered a pattern among working middle-aged male adults — they tended to visit the grocery store only during the weekends or at the end of the day on weekdays, and if they came into the store on a Thursday, they almost always bought beer.

Well, when Manager Mike was armed with these facts, he quickly used this information to maximize beer sales on Thursday evenings by offering discounts, bundles, and specials. Not only was the store owner happy with the increased revenues, but Waldorf Food Co-op’s male customers were happy because they got more of what they wanted, when they wanted it.