Using Intelligence based models for neutral analysis

Tags: #<Tag:0x00007f76fb9bbb48> #<Tag:0x00007f76fb9bb8c8> #<Tag:0x00007f76fb9bb2d8> #<Tag:0x00007f76fb9bad88>

This is an archived article from 2015.

Positioning is not key

The written mediums, especially newspapers, are looking at the digitization by calling it a crisis. Looking at the history of newspapers is revealing, because it shows that the origin of Acta Diurna and tipao are directed communication to the plebs - the commoners. Even today the most successful newspapers are gossip focused; what we call boulevard in Europe. The focus for newspapers for centuries has been to deliver positions to the people - a presentable opinion; into the blue

In a time, where internet users can do the same via the means of digital publication, a position loses its value. That is one of the keys why newspapers are facing a crisis. The transitioning of redaction work from positioning-chain articles to neutrally analyzed and easily readable forecasts and suggestions is struggling. Indicators of this are mono-cultural re-posts of large press agencies, the rise of printed press monopolies and last but not least the shrinking of the newspaper publishing sector. 

Using the MOSAIC model

Facts are not enough for an article, when they aren't analyzed. A model based approach can be helpful. A relevant publication requires timing, which is why you do not build your analysis on demand, regardless of what you publish or who your target audience is.
In the following this is a very basic example of the MOSAIC model, which can be used to qualify publications. It can be used like a cyclic waterfall, in a topic-down fashion to start:

The origin of this is from Cybercrime and Espionage: An Analysis of Subversive Multi-Vector Threats by by Will Gragido and John Pirc. It is meant to enhance the information basis for information security based research. 
Tabloid newspaper articles about this topic are of bad quality and harm the professional field. The model can serve as an example which kind of background structure is required to avoid positioning.


A neutral position does not exist, because even the order of arguments and the phrasing are inherently suggesting a tendency. It's important to leave this in between the lines, because relevant publications should start discussions on the internet and in social medias. Strategically this aids the reputation of the source. Lurid headlines are of short term consequence, and not exclusive to either tabloid or investigative journalism.


Motive models can be Game Theory based. This way it is possible analyze the Motivation aspects. 

The motivation is the payoff. In a more mature modeling application in such a graph the edges (links) can be weighted with probability values. The payoffs are either potential in case this is used for a future forecast, or the graph is used to track a decision tree in reverse. 


OSINT methods can involve working with text, which in IT is called parsing. Simple OSINT tools are article similarity (wording, citation, punctuation...)  to add source relations, and other probable meta-information. Such similarities can occur across multiple sources, which builds a similarity network. In case of a newspaper the origin nodes would be Reuters, APA, DPA or others; the copied and pasted source. Since many articles are published in a digital format, the similarity can be calculated without the need to scan paper [1] [2]. In a way the similarity can be used to assign popularity to a topic, since multiple re-posts suggest interest.

Other tools include automated summary [3], which is useful to reduce the language overhead. Modern publications often reuse templates here, which can be ignored if they do not serve information.


Studying the topic can be enhanced with searching skills. Useful tools include custom search engines, additional to generalized ones like Google. There are [4] multiple online indexes available and the results greatly differ. Additionally it's important to have a growing database of publications synchronized, with RSS or Atom feeds. 

Custom indexes should also include local publications on storage mediums [5]. Often only parts of books are relevant to the active topic, and referencing them is easier with search indexes. This way it is also possible to identify experts to ask for analysis or references.


Correlation can be done based on factors like source, time, similarity etc. The source can be the categorized index, based on a search through collected materials. This way central topics asynchronously emerge. This is a non-reactionary analysis approach, and useful because the analysis should reside as a fundamental baseline. An article needs to transports the key issues of the analysis.

From various of these correlated collections it is also possible to confluence a targeted analysis. Depending on the information analysis chain the event can get multiple angles and dimensions, which could interest the target audience. 

Thinking beyond digital

The approach to digitally publish in the same way, like it was printed 500 years ago is not sufficient to satisfy a modern reader audience. 

Nowadays newspapers have to deliver content, which readers cannot google. Of course a large amount of the audience is satisfied with tabloid press, copied articles and simple positions. However this is also easy to deliver, and does not position a publication so that you can sell it. The digital age of publication is not just about new channels. It's also about a change in the information perception, and the distribution.

A wide-scale voluminous distribution of facts and positions via internet technology can be a source for an analysis, but not for direct re-publication.