What is a Hierarchical Model?

The term Hierarchical Model covers a broad concept spectrum. It often refers to a lot of set ups like Multi-Level models where there are various levels of information or data all related be some larger form.

The Hierarchical model is similar to the Network model; it displays a collection of records in trees, rather then arbitrary graphs.

Here is an example of on type of conventional Hierarchical model:

You can see from the above figure that the supplementing information or details branch out from the main or core topic, creating a “tree” like form. This allows for a visual relationship of each aspect and enables the user to track how the data is related.

There are many other ways to create this type of model, this is one of the simplest and is used the most often.

An example of information you would use the Hierarchical model to record would be the levels within an organization, the information would flow such as:

• An organization has several departments
• Each department has several subdivisions
• Each subdivision has sections

So the Hierarchical model for this scenario would look closely like the one below. As you can see this model is substantially larger, the benefit of the Hierarchical model is that it allows for a continuous growth, though it can take up a lot of room.

With each addition of data a new branch on the “tree” is formed, adding to the information as a whole as well as the size.

Hierarchical models allow for a visual parent/ child relationship between data sets, organizational information, or even mathematics.

The idea for these models is to begin with the smallest details, in the example above that would be the sections.

From the smallest details you would move up (it is often easiest to think of the model as a hierarchy) to the subdivisions, above the subdivisions you find departments, and finally ending at one “parent” the organization.

Once finished you can sit back and view the entire “family” of data and clearly distinguish how it is related.

{qbapagebreak title=Hierarchical Model Usage}

How is the Hierarchical Model used?

The first mainframe database management systems were essentially the birth place of the Hierarchical model.

The hierarchical relationships between varying data made it easier to seek and find specific information.

Though the model is idea for viewing relationships concerning data many applications no longer use the model. Still some are finding that the Hierarchical model is idea for data analysis.

Perhaps the most well known use of the Hierarchical model is the Family Tree, but people began realizing that the model could not only display the relationships between people but also those between mathematics, organizations, departments and their employees and employee skills, the possibilities are endless.

Simply put this type of model displays hierarchies in data starting from one “parent” and branching into other data according to relation to the previous data.

Commonly this structure is used with organizational structures to define the relationship between different data sets.

Normally this contains employees, students, skills, and so forth. Yet we are beginning to see the model used in more professional and meta-data oriented environments such as large organizations, scientific studies, and even financial projects.

Though the Hierarchical model is rarely used some of its few uses include file systems and XML documents.

The tree like structure is idea for relating repeated data, and though it is not currently applied often the model can be applied to many situations.

Issues Related to Hierarchical Models

The Hierarchical model can present some issues while focusing on data analysis. There is the issue of independence of observations, when data is related it tends to share some type of background information linking it together, therefore the data is not entirely independent.

However, most diagnostic methods have need of independence of observations as a key hypothesis for the analysis.

This belief is corrupted in the incident of hierarchical data; such as when ordinary minimum square regressions turn out typical miscalculations that are too small.

Subsequently, this usually results in a greater likelihood of rejection of an unacceptable assumption than if:

(1) a suitable statistical analysis was performed, or 
(2) the data contained within honestly self governs observation.

{qbapagebreak title=Hierarchical Model Structures}

Other Hierarchical Model Structures

Though the tree like structure is perhaps the simplest and also the most desirable form for new users there are other types or structures for this model.

Hierarchy is also structured as an outline or indented list. It can be found in the indented lists of XML documents.

The example, below, presents information similar to those above that we have created but the tree like form is not used in this Hierarchical Modeling but that of indentation.

• ORGANISATION
  • Department 1
    • Subdivision 1
      • Section 1
      • Section 2
      • Section 3
    • Subdivision 2
      • Section 1
      • Section 2
      • Section 3
    • Subdivision 3
      • Section 1
      • Section 2
      • Section 3
  • Department 2
    • Subdivision 1
      • Section 1
      • Section 2
      • Section 2
    • Subdivision 2
      • Section 1
      • Section 2
      • Section 3

One thing you must keep in mind at all times is that no matter what type of structure you use for the model you need to be able to add categories at any time, as well as delete them.

An idea to ensure that this is possible is to use a list view or tree view with expandable and collapsible categories.

You can also use the model in a visual form, something involving a cylinder or pyramid or even a cube, this visual presentation of the data would be most suitable for a presentation of data to a group of professionals.

This form would be better for smaller less detailed levels. There is an example using some of the same information from above but shown more compact below.

There are various structures of the Hierarchical Model; in fact there are many more then those shown here.

The type you use all depends on the data you are using. The methods differ according to whether your data is people related, mathematical related, or just simple statistics.

Review of the Hierarchical Model Facts

1. This model expresses the relationships between information. How they are related and what they are most closely related to.

2. The Hierarchical Model is often thought of as a hierarchy. The idea is to think of your data as a family.

3. The model has many different structures and forms. Each is best used depending on the type of data being recorded, the amount of data being recorded, and who it is being recorded for.

4. Speaking in parent/child terms data can have many children but only one parent.

5. The model begins with core data and branches off into supplementing data or smaller related data.

6. One must remember to start with the smallest detail and work their way up.

If you keep to these simple and compacted guidelines your own Hierarchical Model will be successful, clean, clear, and well built. The point is to present information in a simple and easy to read manner.