In 1978, Charnes, Cooper and Rhodes (1978) introduced Data Envelopment Analysis (DEA). These scientists developed the Data Envelopment Analysis (DEA) model without any limitations on production technology to establish the best practical limit. As a result, Data Envelopment Analysis (DEA)’s methodology is oriented towards frontiers rather than central tendencies. In recent years, this methodology has been used in detail by many domestic and foreign researchers in their scientific studies.

Data Envelopment Analysis (DEA) Method compares the production units, which are assumed to be homogeneous, among themselves. First, the best observation is accepted as the efficiency limit. After that, other observations are evaluated according to their proximity and distance to this most active observation. In other words, the activity limit is not a default situation here; it is an observation. Since the efficiency limit is determined this way, the error term is not used in this method, even indirectly.

However, it is thought that separating the observations representing the borderline (very extreme values ​​close to the border) among the observations will provide a healthier result.

The main feature of Data Envelopment Analysis (DEA) is that it is general. In Data Envelopment Analysis (DEA), there is no assumption for functional form. Therefore, Data Envelopment Analysis (DEA) can be performed under Constant Returns to Scale (CRS) and Variable Returns to Scale (VRS) deductions. With this method, efficiency measurement can be made according to both output–oriented approaches with data input and input–oriented techniques. Among these approaches, the process of obtaining the data output with the minor input use tries to determine how much the number of inputs used in production can be reduced proportionally without reducing the data production amounts. On the other hand, obtaining the maximum output with data input is concerned with how much production quantities can be increased proportionally without changing the data input set. However, both measures give the same results when there is a constant return to scale.

Data Envelopment Analysis and Efficiency Measurement

We carry out various measurements to evaluate the units’ performances we work within different fields such as health, energy, and finance. To compare the teams’ performance, we can determine how successful they are with the available resources by comparing the activities of the units.

Data Envelopment Analysis (DEA) is the primary data analysis technique used to evaluate inter–unit activities.

Data Envelopment Analysis (DEA) is a non–parametric efficiency analysis technique used to evaluate the relative effectiveness of similar Decision–Making Units (DMU) among themselves in the implementation process by using the input and output variables at our disposal.

Notice that, since it is a non–parametric method, there is no distribution requirement (normal distribution, etc.).

As a result of the Data Envelopment Analysis (DEA), an efficiency score for each decision–making unit is calculated. If the value of the efficiency score of the decision–making unit is 1, that unit is active; If it is less than 1, we consider it inactive.

Even if this value is 0.999, the relevant unit is not considered active. Therefore, the activity score of the observation that wants to be a functional unit must also be 1.

Data Envelopment Analysis (DEA) is also a very flexible method for determining inputs and outputs. In the analysis phase, we can use a single work and benefit from more than one output variable.

Ultimately, an efficiency analysis method that allows multiple input and output variables.

Data Envelopment Analysis (DEA) is not the only method used for effectiveness evaluation; however, it is the most popular in the statistical literature due to its ease of use and flexibility. In addition, this technique has different approaches such as Dynamic Data Envelopment Analysis (DEA), Bootstrap Data Envelopment Analysis (DEA), and other modeling techniques such as additive Data Envelopment Analysis (DEA).

We can also use alternative models such as Dynamic Data Envelopment Analysis (DEA) according to the input–output structure of Data Envelopment Analysis (DEA).

Computer software also offers both paid and free software in terms of applicability. For example, we can refer to R–Project, MaxDEA, PIM–DEA, DEAOS to implement Data Envelopment Analysis (DEA).

We can use commercial and licensing software such as MaxDEA and free and open–source software such as R–Project.

Data Envelopment Analysis (DEA) has many different sub–problems in itself. In addition, there are many other models in the literature.