On line Experiments
We offer the opportunity to test some projects in phase of study. Before, It is advisable to take a look to some of the referenced articles (publications) in addition to the bibliographical notes. At this moment, you will be able to try some psychological and psychiatric spanish texts under two scientist corpus which deal about “mental disorders”. First, a small corpus (5335 terms/959 documents to 187 dimensions) based on a guide of "mental disorders". The second is corpus of great size (14685 terms 8682 documents to 230 dimensions) in which you will be able to try, in addition to psicopatology terms, several contents on general psychology ( remember that you will have to wait some seconds)
- You will be able to extract semantic neighbors by means of several ways including the form in which the Predication algorithm is activated. For example, you can begin introducing the term “fobia” with normal cosine.
- For the corrections ways you can try introducing “fobia” with corrected cosine and will extract neighbors with generally greater vector (column of the right “mod”), which means that they are terms that the system “knows better”.
- In order to get an idea of the predicatives structures we recommend you this technical report in [pdf] written for its comprenhension. Even so, you can test “fobia a las tormentas” with the normal cosine and then with both types of predication ways. You will verify that the predication algorithm behaves better for this type of structures (“personalidad de los gérmenes”, “fobia a los perros”, etc)
- You will be able to compare semantically two texts of the subjects that are represented in corpus. You can try introducing two paragraphs that are about psychiatric upheavals. The results will be expressed in cosines (- 1 to the 1)