In the absence of statistics, you need to apply theory or expert opinion. For example, at one time, people were quite interested in using markets for making predictions (see the attached link). In theory, if you created a reward structure where good predictions are financially rewarded, you will get the best predictions. In aggregate, perhaps better than any individual's predictions.

If you have a small amount of data, Bayesian analysis may be a good way to combine expert opinion with the limited data available to make optimal predictions.

My dissertation examined one kind of analysis, comparing empirical and non-empirical methods. I found that empirical methods work really well and non-empirical data has to be very accurate before it's useful. So, I would urge you to try hard to develop machine learning solutions.

https://en.wikipedia.org/wiki/Prediction_market

Dear Alan Mead, Thank you for the answer. So, in your opinion, there are no such better tools than machine learning for the data analysis in cognitive scope? Thanks

I should have mentioned: I did not know what you meant by "cognitive analysis" so I focused on "the purpose of predictions for BI". I thought that there was a good chance that the answer would be helpful regardless of what you meant by "cognitive analysis."

I am so sorry for the ambiguity. What i meant about the cognitive analysis is ab approach to analyze the data (big data) using the cognitive technology, like a cognitive computing. Since machine learning is very popular nowadays for the data analysis and it is very useful to analyze the data.

Do you mean "cognitive analytics"? I think "cognitive analytics" just means "analytics" that can handle nonmetric data (such as written language) and analytics that learn over time. So, the literal answer to your question is that you cannot do cognitive analytics without machine learning. The definition of "cognitive analytics" implies machine learning (or some kind of statistical analysis).

Thank you very much Alan, that is what i have to know recently, to make sure that machine learning is the key for cognitive analytics.

Intinya bagaimana mengekstrak pengetahuan tacit menjadi eksplisit secara otomatis right?? Tanpa machine learning??

untuk tacit extraction mungkin masih beberapa step di depannya Dr. Tb., dan apakah ada tool yang lebih baik dari machine learning untuk cognitive teknologi? 

Dear Prof. Herison Surbakti

It is interesting, I wonder what made you want to investigate a non-statistical approach? 

Big data in my view, does not obey, in contrary to the conventional assumption, to the basic law of large numbers (the law that is responsible for converging the results to  steady figures through repetition). 

Another failure that needs to be addressed is regarding the often neglected first step in machine learning, which is the creation of expert knowledge. Without substantial and relevant expert knowledge to lean on, the machine is indeed learning, but what??!? 

Machine learning is first of all human learning, 

I discuss these issues in my project "Philosophy of Data Science" (GT data mining)

Best luck

Edith

Thank you very much Prof. Edith Ohri 

i am now following your project. very interesting to find my path in analytics and data mining. thank you