Replies to post #162407 on Anavex Life Sciences Corp (AVXL)

[Doc328]

Steady-T is right that the purpose of the KEM analysis was to generate testable hypotheses.

Don't trust me? Well let's see what Ariana says:

"KEM® uses association rules to fully explore complex datasets in order to reveal hidden relationships and to derive new hypotheses" link to Ariana

And what did Afshar and Missling present about this in the past?

"Identi?es most relevant and powerful causal relations, revealing hidden relationships and deriving new hypothesis" link to Avanex

Also,

Dr Missling said, multiple times that the information generated

was Non-hypothesis, data driven

That's not what he said. This is what he said earlier this week from SA transcript

"In order to identify any biomarkers, we have taken a completely unbiased data driven hypothesis free approach to the systematic analysis of our Phase 2a study.

Using the artificial intelligence technology KEM from Ariana Pharma, the relation between all parameters both clinical and genomic and the impact on outcome was ranked in an entirely unbiased way. The use of machine learning rather than a commonly used hypothesis driven method, in which you select a hypothesis, which is then tested against data, has several advantages. The machine learning methods are objective and purely data driven. They enable the analysis of very small cohorts of patients, which systematically ranks biomarkers without any bias. And explores the discrimination and patient effect. That means there is no need for an prior hypothesis. So, it is more like just let the pure data speak."


So by that statement, he is saying they go into the analysis without a hypothesis. So they don't go in saying ApoE matters or SigmaR1 matters they just try to find the strongest associations. I would (and I think Steady would) agree with Missling and his statement (though convoluted) is in agreement with Ariana and Anavex's prior slideset. Then the analysis is done and the analysis produces a hypothesis that can be tested with a new cohort (the Ph2/3) see slide 15 in the above Anavex slideset.