| title |
|---|
Week 7 |
HOME | Timeline | Previous Week | Next Week
From the result of conditioned on 1 variable, it can be inferred that
- price is independent of beds always.
- price is independent of types given any other variable.
- price is independent of sqfeet, but when conditioned on baths it is dependent.
After that, I tried to run the conditional dependency test on 2 variables for inferencing the causal model by eliminating the dependencies. But the test was taking too long. After waiting for 6.15 hours, I aborted the test.
I ran causal discovery tests, and it is giving different results on different executions. And the following were the 2 results that were more frequent than the others. One important result is that price is the only exogenous variable.
While working with gridBundle, I choose to use some snippets from housingMain for debugging purposes. Then I found a bug and error in the housingMain code. I have fixed the bug, and now I can see a change in the causal model.
The resulting causal model can be seen as follows. And in this beds is an exogenous variable.
While trying to debug the housingModel in python, I observed that isIndependent test in python in because module is resulting in TypeError: 'NoneType' object is not iterable.
Roger noticed some irregularities in the dataset and found that the dataset is not aligning with the real-world cases like the average prices of the house from the dataset are far different from reality. So I was asked to find another dataset that can be used for the analysis.
In the Team meeting from 1230-1430, Roger helped me with the gridBundle. I have also reported the bug in isIndependent test in the because module.
I was looking for any datasets where some reasonable hypotheses can be formed. After looking over more than 100 datasets in Kaggle and trying to make my dataset in WorldBank.org, I found a few datasets, where some real-world hypotheses can be formed, but I think they are not intuitive enough.
HOME | Timeline | Previous Week | Next Week