Co-analysis with numerical data such as gene expression information is also possible
AI that realizes the utilization of medical data
Quantify and analyze natural sentences in the healthcare field with "vectorization of words and documents".
When reading a sentence, understand not only the meaning of the word alone, but also the meaning and position of the words that appear around you.
You can analyze the composition of morphemes of sentences in a document at once.
Realize EBM (evidence-based medicine)
In the medical and healthcare arena, the right information must be carefully selected, as actions taken based on unfounded and incorrect information can sometimes have serious health disadvantages. ..That is why EBM (Evidence-based Medicine) is so important in the field of healthcare. Concept Encoder emphasizes EBM and does not black-box the process of producing results, so it has "objectivity" (which other people can understand and understand), "transparency" (the process of producing results is clear), and Achieve "reproducibility" (same result even if others do it).
Learn co-occurrence of words in a document and vectorize
Concept Encoder is a method of morphological analysis, that is, an analysis method that evaluates the frequency of appearance of morphemes (the smallest unit that has meaning in a language) among multiple documents, using a method called "vectorization of words and documents". The features of the text are quantified.
When a word appears in a sentence, the frequent appearance of another limited word in the sentence is called "co-occurrence", and vectorization is numerical by expressing this co-occurrence relationship in a matrix. Perform analysis.In other words, by numerically expressing and clarifying co-occurrence relationships, it becomes possible to find differences in relevance and importance without defining the meaning of words.
Analyzing the co-occurrence / composition of words and the composition of morphemes of sentences in a document
The Concept Encoder starts by learning the co-occurrence of words in a document and vectorizing it.In normal vector analysis, the method of finding the sentence and the word separately and then finding the relevance may be taken, but in Concept Encoder, the co-occurrence / composition of the word and the morpheme of the sentence in the document You can analyze the configuration collectively. FRONTEO is patented for this technology.Finally, we will learn and optimize.
Tag the documents you want to find and learn what's important
By tagging the documents you want to find and letting the Concept Encoder learn what's important, you'll optimize that document for a higher score and higher ranking.