Sentence Analyzer - enabling business and enterprise applications to handle sentences and text |
| Product Reviews - measuring customer delight Product reviews invite analysis, sentiment mapping. But before that, "hearing" the right clues may be a good idea. |
Sentiment mapping begins with a sentence, not a "bag of words". This demo suggests a marketing experience driven sentiment capture.
|
| Experienced sales and marketing people can "tell". They listen for key phrases in customer feedback. These key phrases are very specific to the product, customer demographics, and take years of selling experience to recognize. This toolkit has provisions (not in this demo) for weighting the entire sentence on occurence of positive or negative sentiments, cancelling things out, flavoring them strong or weak etc. But firstly, a marketing person can segregate feedback sentences as "complete delight", or "qualified brickbat", according to the vibes of the times. This way, you can incorporate the tone and nature of a sentence as part of the analysis. Sentences (and not math based on words per se) are the essence of the customers feedback. |
| In the first text, there is a bit too much of "good". Please also try second "bad" review text. Please note that this demo does not include "negativizing" or "coloring" or "flavoring" a sentence, although such things exist in the toolkit. |
| Note: Creating benchmarks for unstructured text needs some pre-processing of a biggish set of typical reviews, as well as some mechanical picking of the best fit benchmarks. |
| One major feature suggested by this demo, as well as the one in text2data.net, is that any extraction tool must have business people configuring the extraction, to some very specific knowledge that they alone have, and programmers ought not to go the last mile. |
| Try out the below , and sometimes, explore the "Compare less tightly" and other options. For questions : kinshuk_in @ yahoo dot. com
|
| Note : The RESULTS shown are the human readable equivalent of Java/C# objects, and they have lots of additional information, like word meanings, group codes like colors and flavors etc., intended for further analytical/statistical treatment. This demo (created with no NLP APIs), stresses that with text, it is better to first maximize grammar based processing, and use statistics/math methods much later. |
| Sentences must be separated from each other by an ending period (. or ! or ?) and one space. | Skip the descriptive stuff and go directly to demo |
| Product Reviews : getting user feedback ( Click more samples :
then Find/Collect)
|
| Enter DESIRED benchmark sentence(s), or click one of the sample buttons above.. Max. 5 sentences. |
|
|
| Enter TARGETed sentence(s) or a paragraph. [See a Product user review extract below]. This is what you want compared against the benchmark(s). Max. 20 sentences. |
|
|
|
|
(Please scroll down for the RESULTS)
|
| Summary results : Searched and sorted among 8 target sentences. Highest co-relation 50.0 percent. |
| Best matching content + structure after a find/collect operation |
|
|
| All sorted matches (Best finds on top, degrades towards the end, and very bad matches ignored) |
|
|
|
|
- Features, tall claims, and things to note in this demo :
- .
|
|
|
| More view/try pages here. |
|
|
More reading for those interested... |
| 1. Things that are not obvious from the demo |
| 2. Business products and possibilities |
| 3. So what !! Universal grammar has been in use for decades now ... |
| 4. The inevitable comparisons, to what already exists out there. |
| 5. What is a sentence, to future application builders ? |
| 6. The genesis and design principles story |
| 7. Arbitrary listing of business usages |
| 8. Extensions, additions, customizations possible in the toolkit |
| 9. Important : Combining with the document extraction tool at text2data.net, benefits |
| Sometimes the math attempts to become the marketing. Trust sentences, because they are more human than numbers. |
| Contact at : kinshuk_in @ yahoo dot. com |