Classification model
The classification model is in some ways the simplest of the many types of predictive analytics models we are going to cover. It places data into categories based on what it learns from historical data.
Classification models are best for answering yes or no questions, as they provide a broad analysis that is useful in guiding a decisive action. These models can answer questions such as:
For a reseller: “Is this customer about to resign?”
For a loan provider: “Will this loan be approved?” or “Is this applicant likely to be in default?”
For an online banking provider: “Is this a fraudulent transaction?”
The breadth of possibilities with the classification model — and the ease with which it can be retrained with new data — means that it can be used for many different industries.
Cluster model
The cluster model sorts data into separate, embedded smart groups based on similar attributes. If an e-commerce shoe company seeks to implement targeted marketing campaigns for their customers, they can go through hundreds of thousands of records to create a tailor-made strategy for each individual person. But is this the most efficient use of time? Probably not. Using the cluster model, they can quickly separate customers into similar groups based on common characteristics and devise strategies for each group on a larger scale.
Other uses of this predictive modeling technique may include grouping loan applicants into “smart buckets” based on loan characteristics, identifying areas in a city with a large amount of crime, and benchmarking SaaS customer data into groups to identify global usage patterns.
Prediction model
One of the most widely used predictive analysis models, the forecasting model deals with metric value prediction and estimates a numerical value for new data based on experience from historical data.
This model can be used wherever historical numerical data is available. Scenarios include:
A SaaS company can estimate how many customers they are likely to convert within a given week.
A call center can predict how many support calls they will receive per hour.
A shoe store can calculate how much stock they need to have on hand to meet demand during a particular sales period.
The forecast model also takes into account several input parameters. If a restaurant owner wants to predict the number of customers she is likely to receive in the following week, the model will take into account factors that may influence this, such as: Is there an event nearby? What is the weather forecast? Is there a disease on the way?
Outliers model
The Outliers model is oriented around abnormal data entries in a data set. It can identify abnormal figures either alone or in conjunction with other numbers and categories.
Detecting an increase in support calls, which may indicate a production error that could lead to a recall
Finding abnormal data in transactions or in insurance claims to identify fraud
Finding unusual information in your NetOps logs and noticing the signs of impending unplanned downtime
The Outlier model is particularly useful for predictive analytics in retail and finance. For example, when fraudulent transactions are identified, the model can assess not only the amount, but also the location, time, purchase history, and nature of purchase (i.e., a $ 1000 purchase on electronics is not as likely to be fraudulent as a purchase of the same amount on books or ordinary aids).
Time series model
The time series model includes a sequence of data points captured using time as the input parameter. It uses last year’s data to develop a numeric metric and predicts the next three to six weeks of data using that metric. Use cases for this model include the number of daily calls received in the last three months, sales for the last 20 quarters, or the number of patients who showed up at a given hospital within the last six weeks. It is a potent means of understanding the way a single metric evolves over time with a level of accuracy beyond simple averages. It also takes into account seasons or events that may affect the metric.
If the owner of a salon wants to predict how many people are likely to visit his business, he can turn to the rough method of calculating an average of the total number of visitors over the past 90 days. However, growth is not always static or linear, and the time series model can better model exponential growth and better adapt the model to a company’s trend. It can also forecast for multiple projects or multiple regions at the same time instead of just one at a time.
