> For the complete documentation index, see [llms.txt](https://trusta-labs.gitbook.io/trustaai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://trusta-labs.gitbook.io/trustaai/products/trustgo/media-scoring-methodology.md).

# MEDIA Scoring Methodology

### ***Computational Logic***

<figure><img src="https://2950890397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FQZJLzGNbGstalGZITcdI%2Fuploads%2F3WwpimMG5w4n81ZHnicg%2Fimage.png?alt=media&amp;token=f54d6458-05f3-4aa4-8ad7-2b752d2eecfc" alt=""><figcaption></figcaption></figure>

The Diagram is an illustration of the bottom-up computational logic of the MEDIA score:

1. The first step is to transform and normalize the variables using a normalized tunable sigmoid function. This function ensures that the values of the variables are mapped to a standardized range, allowing for consistent comparison and analysis.
2. For each of the five dimensions (Monetary, Engagement, Diversity, Identity, and Age), a subscore is computed. This is achieved by taking a weighted sum of all the variables within that dimension. Each variable is assigned a weight that reflects its relative importance in determining the overall score for that dimension.
3. After calculating the sub-scores for each dimension, they are scaled to a range of 0 to 100. This scaling process standardizes the sub-scores, making them easier to interpret and compare across different dimensions.
4. Finally, the MEDIA score is calculated by taking a weighted sum of all the sub-scores. Each sub-score is multiplied by its respective weight, reflecting its significance in contributing to the overall value assessment.

### ***Sigmoid Transformation***

![](https://2950890397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FQZJLzGNbGstalGZITcdI%2Fuploads%2F1MHG8kAcBCMkBZZWoDSh%2Fimage.png?alt=media\&token=c62ba2e1-ac66-4aa8-bdd1-5ac57183ede3)![](https://2950890397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FQZJLzGNbGstalGZITcdI%2Fuploads%2F3vvXyp4K2RYPZfCCEpRi%2Fimage.png?alt=media\&token=18495d2f-baa4-4651-a8a0-40d89f7e964a)

The Sigmoid function， represented by the equation $$y = \frac1{1+e^{-x}}$$ , is a non-linear S-shaped transformation function. As the input values x increase, the output y gradually transitions from 0 to 1. This gradual transition allows the sigmoid function to capture non-linear relationships. In the normalized tunable sigmoid function used in the MEDIA scoring system, there is a parameter that controls the pace or speed at which y transitions from 0 to 1 as x increases. This parameter allows for fine-tuning the behavior of the sigmoid function to match the desired range and sensitivity of the scoring system.