Hello
I'm working on a research on The effect of certain demographic information, Say population density, on people's response at different stage of Covid 19 pandemic.
Now the problem is, the independent variable, population density, does not change over the Short time period In this study ( two months). Our dependent variable, People's of social distancing behaviors, On the other hand, Change at different period. My question is how can we deal with this kinds of regression? we are thinking about divide the pandemic into three period: early stage, peak, decline stage. We will regress population density on people's reaction at different stage separately. So here I will have three separate regressions. I'm wondering if this Approach could cause any problems, Since this is not a typical time series Model ( with the independent variable not changing overtime).
Thank you for any suggestion.
I think that you can work first in discrete time, call it t.
The time unit may be one day, or one week, or one month: your choice. This is the time between t and t+1...
In models I will share with you, I have used typically the hour or the day as a unit. No need to decide in a rigid way at this stage.
Then you need to define your agents, and go for agent based models: a human is modeled best in that way.
Call an agent (person) x, and your population set X, of all agents considered.
What is your concern?
Social distancing behaviour, maybe. Then define r(x,t), the perceived safety sphere radius of x at time t. Simply said: if people are closer than r to x, x is not happy and tries to avoid this, while at distance above r, x is happy.
If agent x comes across agent y, at t: d(x,y,t) is the parameter monitored. Is it above r(x,t)? Is it above r(y,t)? Then fine, otherwise an avoidance strategy (walking away or around, etc) by the one of the two x or y who feels most at risk: z such that r(z,t) = max (r(x,t), r(y,t)), z wants more space...
All the above is straightforward to programme in any computer language.
You have the start of your story, your simulation.
Now you want to assess the covid19 transmission risk, hence you want to test the probability that x stays healthy after N interactions with different agents y1, ..., yN.
See my model which was developed for close interactions (meals at a table, typically). You can go from there and develop one suitable for you.
Preprint Covid19-transmission-model-Elderly Care Home-etc-2020-04-28
Article Covid-19: Modelling propagation with Agent Based Models (ABM...
Let me know if you have more details on what you want to do with your modeling, and data sets...
I hope this helps you
I think that you can work first in discrete time, call it t.
The time unit may be one day, or one week, or one month: your choice. This is the time between t and t+1...
In models I will share with you, I have used typically the hour or the day as a unit. No need to decide in a rigid way at this stage.
Then you need to define your agents, and go for agent based models: a human is modeled best in that way.
Call an agent (person) x, and your population set X, of all agents considered.
What is your concern?
Social distancing behaviour, maybe. Then define r(x,t), the perceived safety sphere radius of x at time t. Simply said: if people are closer than r to x, x is not happy and tries to avoid this, while at distance above r, x is happy.
If agent x comes across agent y, at t: d(x,y,t) is the parameter monitored. Is it above r(x,t)? Is it above r(y,t)? Then fine, otherwise an avoidance strategy (walking away or around, etc) by the one of the two x or y who feels most at risk: z such that r(z,t) = max (r(x,t), r(y,t)), z wants more space...
All the above is straightforward to programme in any computer language.
You have the start of your story, your simulation.
Now you want to assess the covid19 transmission risk, hence you want to test the probability that x stays healthy after N interactions with different agents y1, ..., yN.
See my model which was developed for close interactions (meals at a table, typically). You can go from there and develop one suitable for you.
Preprint Covid19-transmission-model-Elderly Care Home-etc-2020-04-28
Article Covid-19: Modelling propagation with Agent Based Models (ABM...
Let me know if you have more details on what you want to do with your modeling, and data sets...
I hope this helps you
We can look for a clean and population growth factor for study, for example
Population growth
The world population in 1950, five years after the creation of the United Nations, was estimated at approximately 2.6 billion. According to United Nations estimates, the population reached 5 billion on 11 July 1987, and the number reached 6 billion on 12 October 1999. In October 2011, the world population reached 7 billion. To celebrate the occasion, a global movement called "7 Billion Actions" was launched. The world’s population is expected to increase by two billion people in the next 30 years. This means that the world’s population will increase from 7.7 billion today to 9.7 billion by 2050, and that the number will reach 11 billion by the year 2100.
The factor in this significant growth is the increase in the number of individuals of reproductive age, accompanied by significant changes in fertility rates, increased urbanization and an acceleration of migration. These trends have far-reaching implications for future generations
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Regression