Viewing 6 posts - 1 through 6 (of 6 total)
  • Author
  • #43384

    Hello. Let me first apologize if this is already answered elsewhere.  I have been collecting spectra from single origin chocolate with the hope of being able to determine provenance using the SCiO.  You can see progress in the image attached (I need more samples).


    My question is this.  Is it possible to create a model that essentially does a linear combination of categorical data, a model created from a set of single origin bars that can correctly identify, for example, a chocolate bar that is 50% Madagascar and 50% Venezuela?




    You must be logged in to view attached files.



    I apologize for the delayed response.


    Your suggestion would be feasible providing that the spectral fingerprint of chocolate bars of different provenances are different enough from each other. 


    As we have not tested it by ourselves, further research is needed.

    Our chocolate applet which is now available as part of our SCiO consumer app, provides you with nutritional information and cacao levels estimation in chocolate samples.






    Thank you,  Ayelet.


    I thought I understood your response, but I am not sure.  Using your Pain Reliever model, which accurately identified an Acetominophen tablet and an Ibuprophen tablet, I crushed and mixed the two.  The app returned a ” Material not recognized” result rather than a blend.


    I realize my chocolate model has weaknesses, and have not blended, say Guatemalan and Peruvian chocolates to see, but I suspect I’d get the same ‘ not recognized’  response.  Or do you think if the model were strong it would give me a fractional blend value?


    Thank you very much,



    I think I am asking a more general question, using specifics as examples.  For instance, If I sampled flours – wheat, rye, millet, buckwheat,…. and created a good categorical model that could identify each.  If I then mixed two different flours together, how can I create a model that will say ” 50% rye, 50% millet”?  I think right now the models just try to find the best categorical fit and does not try to match linear combinations of categories, right?







    I have not heard back on this – so the question is can SCiO modeling determine linear combinations of categorical data?  How can  I determine fractions in a blend without mixing many different combinations?





    Hi Jonathan,


    I have answered your question in the new post.



Viewing 6 posts - 1 through 6 (of 6 total)
  • You must be logged in to reply to this topic.