Algorithm

An algorithm is a fixed computational process for solving a problem.

Definition

Algorithms generally solve problems. Most of the time, these are mathematical problems that are solved in clearly specified steps. It is therefore a calculation process that works according to a certain (repetitive) scheme.

Algorithms in everyday life

Since the heyday of Facebook, I’m sure everyone has heard of the term. In our daily lives, we constantly encounter algorithms that influence our behavior.A traffic light, a navigation device, the search engines and social networks on the Internet; They all work with algorithms.They not only influence us, they control us! So you want to know how an algorithm works? Then you’ve come to the right place!

For example, that’s why Facebook doesn ‘t show you all my friends’ posts, but only the ones the Facebook algorithm thinks are relevant.

It filters these posts based on our interactions with certain people. The same happens, for example. with ads you click on the Internet.

Features of an algorithm

1. Unambiguity: Each instruction must be unambiguous to process.
   In principle, this means that it must be clear what is to be done.
   
2. Generality: An algorithm must solve a large number of problems of the same type.
3. Executability: Each instruction must be executable by a human.
4. Finiteness: The instructions must befinite and the number of resources used must be equal.
5. Determinacy: The same input always produces the same results.
6. Termination: The process is completed after a finite number of steps and delivers a result.

The technology behind algorithms

As with Google, there are often a variety of algorithms available and consideration must always be given as to which is most appropriate, as runtime efficiency plays a large role.

This means in general: how many steps does an algorithm need to reach a result? For this purpose, an evaluation of the algorithms takes place.

In connection with search engine optimization (SEO), it is important to have heard the terms PageRank and HITS algorithm, as these form the foundation for the algorithms.

Also important in this context are the terms determinism and termination.

Determinate

If an algorithm has only one way to proceed at any time, it is called deterministic. The same input always produces the same result and the steps are clearly defined. If there are several possibilities between which a decision can be made freely, the algorithm is non-deterministic.

Scheduling

However, if the algorithm does not find a way to continue, it must terminate. One speaks of termination when the algorithm reaches a result after finite steps. There is no upper limit to the number of work steps.

Relevance for SEO

PageRank is an algorithm that comes from Google and evaluates web pages with a value (integers only) between 0 and 10.

The basis for this is the linking structure (read backlinks). There are also algorithms that rate the content of a page. According to Google, there are over 200 factors (which only Google itself knows in detail) that play a role in the ranking of a website. Other important algorithms are for example the already mentioned HITS (Hypertext induced topic selection), the TrustRank and the Hiltop algorithm. These algorithms are essential for the ranking of the search engine results lists(SERP).

Strictly speaking, these consist of various algorithms that use numerous signals to display a web page higher or lower in the search results. According to Google, more than 200 different factors are used in weighting web pages. In addition to links and terms, this also includes the topicality of the content of a website as well as its location.

However, the exact formulas, structures and implementations of such algorithms are top secret.

Conclusion

Algorithms facilitate our everyday life and can analyze our behavior. They control and influence us, filter and are thereby intransparent. Algorithms can now be found everywhere and work independently. All search engines known today (also e.g. job search engines) could not exist without a self-learning algorithm.