Here are my latest ruminations on this and that, what is and what is not.
When creating a model that mimics a phenomenon perfectly, one is tempted to interpret the model as a full explanation of it. This is errant causality. It is the profound weakness of modeling.
It is a never good idea to invite administrative jobs in the parlor door. They promise more than they deliver. They ask more than they give. It should be no other way. The converse, to give more than they ask, renders as undesirable the newly appointed administrator.
In the pursuit of truth, someone is always threatened.
We are more likely to accept what we want to believe, or what fits within our precepts, or supports our goals - all regardless of the facts of the matter. Rejection of what we wish to be so is difficult.
We live in a big data world but we still have small data brains. This is not to imply the brain has a small capacity, but it thinks in small data settings. It uses tools such as instinct, intuition to filter or distill large amounts of information to that manageable. The human brain can analyze a maximum of 5-6 variables. But it can’t analyze the value of 50 or more variables or dimensions. Moreover, it can't analyze them in combination, in correlation. But this is today's situation. What we can see are but a few slices or projections of these many dimensions, and this is what it relies on to make decisions and predictions.
Indeed, a great deal of what is call data compression is a recognition a few key data dimensions from which the original data can be reasonably reconstructed. So, data compression is truly a human inspired effort, though applied from of the necessity of limited data capacity and bandwidth.
Subjective vs Objective. What's the difference. statistically speaking?
Statistics is an excellent tool, honest and reliable. One the one hand, it illustrates its weakness by its use of observed data, and on the other hand it illustrates its strength on the basis of the underlying assumptions of the data type and consequent probabilities. Thus, one can sometimes argue either side of the point.
So, what is subjective and what is objective in all this? For the later, it is the probabilistic (i.e. statistical) assumptions about the data and the volume of data. For the former, it seems to be the null hypothesis, as this implies a confirmation or denial of some proposition. Regardless, there is a strictly objective decision.
Despair is a condition brought on by what appears to be impossible, whether an event, goal, or object.
When creating a model that mimics a phenomenon perfectly, one is tempted to interpret the model as a full explanation of it. This is errant causality. It is the profound weakness of modeling.
It is a never good idea to invite administrative jobs in the parlor door. They promise more than they deliver. They ask more than they give. It should be no other way. The converse, to give more than they ask, renders as undesirable the newly appointed administrator.
In the pursuit of truth, someone is always threatened.
We are more likely to accept what we want to believe, or what fits within our precepts, or supports our goals - all regardless of the facts of the matter. Rejection of what we wish to be so is difficult.
We live in a big data world but we still have small data brains. This is not to imply the brain has a small capacity, but it thinks in small data settings. It uses tools such as instinct, intuition to filter or distill large amounts of information to that manageable. The human brain can analyze a maximum of 5-6 variables. But it can’t analyze the value of 50 or more variables or dimensions. Moreover, it can't analyze them in combination, in correlation. But this is today's situation. What we can see are but a few slices or projections of these many dimensions, and this is what it relies on to make decisions and predictions.
Indeed, a great deal of what is call data compression is a recognition a few key data dimensions from which the original data can be reasonably reconstructed. So, data compression is truly a human inspired effort, though applied from of the necessity of limited data capacity and bandwidth.
Subjective vs Objective. What's the difference. statistically speaking?
Statistics is an excellent tool, honest and reliable. One the one hand, it illustrates its weakness by its use of observed data, and on the other hand it illustrates its strength on the basis of the underlying assumptions of the data type and consequent probabilities. Thus, one can sometimes argue either side of the point.
So, what is subjective and what is objective in all this? For the later, it is the probabilistic (i.e. statistical) assumptions about the data and the volume of data. For the former, it seems to be the null hypothesis, as this implies a confirmation or denial of some proposition. Regardless, there is a strictly objective decision.
Despair is a condition brought on by what appears to be impossible, whether an event, goal, or object.
Comments
Post a Comment
Please Comment.