Generating Filtering Rules
A rule filter is an object that contains a collection of filters and constraints. These filters and constraints restrict the set of valid rules or sequences in an Association rules mining task or in a Sequence rules mining task.
You can use rule filters to specify data to build mining models.
Rule filters can contain the following classes of constraints:
Restrict the maximum allowed number N of rules in the model to be generated. If more than N rules are found in the data, the count constraint formulates a priority criterion that selects the N best rules.
Specify the items or the categories of items that must be included or excluded in particular parts of the rules. For example, items that must be included in the rule body or in the rule head of the rule.
Restrict the valid range of particular statistical or numeric properties of the rules. For example, range constraints can restrict the maximum rule length, the confidence, or the support of the rules.
Data mining is used in the following fields of the Corporate Sector:
Finance Planning and Asset Evaluation: It involves cash flow analysis and prediction, contingent claim analysis to evaluate assets.
Resource Planning: It involves summarizing and comparing the resources and spending.
Competition: It involves monitoring competitors and market directions.
The definition of customer profiling is “a description or analysis of a typical or ideal customer for one’s business” Harper Collins Publishers. Customer profiling is a marketing tool that businesses use to understand their customers and helps to make better business decisions. Profiling results in customer profiles which describe customers based on a set of attributes.
Customer profiling helps to understand customers, highlighting who they are, what they look like, what their interests and wants are, etc. This insight recognizes the customer’s characteristics and traits. Having a better understanding of customers helps the company to communicate with the target customers more effectively.
Analyzing transaction data can identify patterns and trends among customers by segmenting the needs from the profiled customers. As a result, the customer profile is not only based on demographic information but also takes into account the past behavior. Analyzing of a customer profile is by the product type, value, frequency, and patterns of spending. It provides a clear picture of most profitable customers by analyzing the different patterns of behavior or spending by customers. Next step after understanding the customers’ type is to develop a strategy allocating the resource for each segment group and optimizing marketing budget and improving return on investment.
Customer profiling models
Profiling is typically reached by funnel groups and leak out customers according to similar features, providing a segment of customers. Collecting demographic, social, and behavioral data make it possible to classify and measure customer or potential customers according to the different social categories, habits, incomes, living standards, preference, age, wealth, and location. These features can be used in the segmentation and drain customers.
Behavioral classifications give the best result in the segmentation of customers because it is based on the customer’s general behavior, likes, and interests. Behavioral classification provides a more in-depth and more reliable look at the customer base and what drives them and motivates them when making purchasing decisions. MOSAIC, ACORN, and Personicx are all well-known examples of the classifications of profiling behavior and hobbies, as well as demographic and social variables.
If customer segmentation powerful tool is used in all customer lifecycle, at least double increase in the sales can be reached. The advantage of customer segmentation is that they can be targeted at each customer segment using an individual approach, rather than one size suitable for all technologies, focusing on a specific segment facilitates communication with customers using a message related to them, and providing a more personalized approach with appropriate marketing communications.
The most common forms of customer segmentation are:
- Geographic segmentation: Considered as the first step to the international marketing, followed by the demographic and psychographic segmentation.
- Demographic segmentation: Based on variables such as age, sex, generation, religion, and occupation and education level.
- Firmographic: Based on the features such as the company size (either regarding revenue or number of employees), industry sector or location (country and region).
- Behavioral segmentation: Based on the knowledge of, attitude toward, usage rate, response, loyalty status, and readiness stage to a product.
- Psychographic Segmentation: Based on the study of activities, interests, and opinions (AIOs) of customers.
- Occasional Segmentation: Based on the analysis of occasions (such as being thirsty).
- Segmentation by benefits: Based on RFM, CLV, and others.
- Cultural segmentation: Based on the cultural origin.
- Multi-variable segmentation: Based on the combination of several techniques.
Customer segmentation techniques are:
- Single discrete variable (CLV, RFM, and CHURN)
- Clustering: K-means, hierarchical
- Latent class analysis (LCA)
- Finite mixture modeling (e.g., Gaussian mixture modeling)
- Self-organizing maps
- Topological data analysis
- Spectral embedding
- Locally-linear embedding (LLE)
- Hessian LLE
- Local tangent space alignment (LTSA)
- Random forests, decision trees
- Deep learning