Analysis Method

Let us introduce you how Neticle Text Analysis works.

We have developed our own searching and analysis methods to get to know any text's meaning without having to read it.

in my experience, this is a bad car

Web Opinion Index: -2

Recognized labels: car

"keyword_stats": {
"total_keyword_hit_number": 1,
"total_synonym_hit_numbers": [
"car": 1

The system is keyword based which means you can to add keywords and their synonyms. It will find every text that contains the given word and analyses it based on the keyword's context.

The system analyses the hits automatically. Our system rate them between -3 and +3 based on the strength of the polarity of the keyword. We named this opinion index.

The scores of the keyword are added up, so you can see the overall opinion and judgement of the given keywords.

We can manage negations and double negations and we can also recognize irony.

We can customize the algorithm if needed (e.g.: if the field or industry of a given keyword requires specific expressions to recognize in the contents).

"recognized_negative_phrases" : [
"related_pos_phrases" : [

"phrase": "a bad",
"mention_number": 1,
"related_entities": [

"mentions": [
" this is a bad car"
"related_neg_phrases": [


This is how we have created our market leader algorithm which is able to analyse the texts near human-level precision.

If two people manually score thousands of texts based on their sentiment, their opinion match at the maximum of 82% because of human subjectivity and others. Neticle's algorithm performs above that generally.

Besides sentiment, we also recognize what the hits are about. Our algorithm automatically recognizes places, topics, persons and brands.