All Cikisi’s solutions have powerful search capability, which is essential for any organization.
Searching information about a particular subject. It often consists in a quick search on a period range to check or complete existing information.
Search and aggregate information about a theme to study trends, details, history. Often used in market studies, state-of-the-art studies, etc.
The most relevant results from your search are presented first.
Relevance is calculated based on:
Occurrence: The more the query keywords appear in an article, the more the article is relevant.
The ratio of the number of articles in our index and the number of articles containing the keyword (idf: inverse document frequency):
Thus, common words like “and” or “the” present in many articles will bring less relevance than a keyword such as “constitution”.
With artificial intelligence, items are sorted based on your previous quotes.
The results are sorted by date, either from the most recent or the oldest article, allowing you to learn about facts over time.
Articles with the most Facebook engagement actions (likes, shares, comments) are shown first. This feature is particularly useful for detecting buzzes, consumer and voter interests, etc.
Results are displayed based on the average ratings, assigned by other readers in your organization. A way to focus on articles with higher quality and interest.
This option allows you to group duplicates within your results.
From a specific date to a specific date (from … to …), or matching a range period (last hour, day, week, month, year).
Country of origin of the source
Source categories. Some sources are categorized manually by our “news” team, other “technologies”, “robotics”, “health”, etc. You can also categorize your own sources according to your needs.
Concept: the content of the article must meet a concept (predefined and stored boolean queries) that you have created
Automatic classification: the content of the article must have been classified by Cikisi (IPTC Media Topics).
Geographical location: search for items located around a given geographic coordinate.
Smart Data allows you to explore your data using our explore module.
All of Cikisi’s articles are subject to semantic enrichment. This enrichment is particularly effective in more than a dozen languages. The platform automatically detects and extracts the following named entities: location, person and organizations (brands, companies, organizations). Semantic enrichment can also annotate articles with keywords, those who strongly qualify the article, generally made of common names.
All articles on Cikisi’s platform are automatically classified with a tag. Do not confuse the automatic classification made to the content of the article with the categories of sources, which are generally assigned manually. Example: An article from a newspaper will be manually categorized as “news” (because the source has been initially qualified as news) but if the article is about a football match, the Cikisi platform will automatically assign the article with “sport” and “football” tags.
The locations identified in the content are converted into geographical coordinates and are completed with those extracted directly from the other elements of the articles (image EXIF file, XML file, etc.).
For every result of your query, Cikisi’s platform will provide you with similar articles and will also recommend articles relevant to you using the artificial intelligence module*.
* Starting with version Cikisi WMT 2.0
Within the SEARCH module of our web platform “Cikisi WMT”.
From your applications (such as an intranet portal), using Cikisi WMT API.
Build your corporate search engine so every member of your organization can access their articles of interest
With a simplified version inside Cikisi Mobile WS.
Unlike Google, which is essentially dedicated to consumers, Cikisi’s platform has a search capability developed specifically for businesses and information professionals.
Paid ads (Google Adwords), Google-oriented SEO strategy (mobile friendlyness, backlinks, metadata, etc.). At Cikisi, the sorting criteria are transparent and are stricltly based on the content, not its form. Even if a relevant article is poorly referenced on a web site with poor mobile optimization, it will be properly promoted on Cikisi’s platform!
Google resultst will be promoted according to your location at the time of your seatch and therefore may vary over time. All information professionals know that a powerful company search engine should not be tuned like a consumer-oriented engine. Cikisi will return and sort articles that exactly match your requests, regardless of whether they come from a French or Japanese source. You may restrict the language of the returnerd articles if you want to.
Travel or shopping related documents are generally not relevant to your business work. Therefore, they can be excluded from indexing processes on Cikisi’s platform.
For example, as of April 26, the “trump AND mexico” search returned 142 results on Google’s search engine. As most Google searches return a maximum of 15 pages (or 150 documents), you need to constantly refine your query to go over this artificial limit. By contrast, Cikisi people are aware that, when running investigation studies, it is essential to return all possible documents to the user. It may represent several hundreds or thousands of documents that must be available within the exploration module. This is why you can access all of Cikisi’s documents when running a search.
As a result, you can forget your privacy but also to the confidentiality of your queries. As Google uses your queries (on a time range and based on locations) to optimize the SEO process or the sorting algorithm, we already demonstrated that it was possible to deduct subjects of interest of some companies. For instance, we were able to demonstrate that company X was interested in a company Y during a merger/acquisition period. Cikisi keeps your searches confidential and does not track or use them for any purpose.
Using images and named entities are very useful for that purpose.
By specifying the set of sources within which you wish to carry out your queries, you can configure a domain-specific search engine, such as health, defense, innovation, insurance, etc. Thus, it enhances the general relevance and reduces the potential noise within results.
Cikisi results can be displayed within Cikisi WMT using the EXPLORE module or within your own applications thanks to the metadata attached to the articles.
Improve your organization's intelligence capabilities with our suite of solutions and our team of experts