Open philusnarh opened 5 years ago
This is a great project
Any reason you're comparing with SA?
The economy of SA is quite stable as compared to Ghnan. So I wanted to compare the two and see how good their respective predictions would be. Perhaps we can conclude that AI is good in predicting for countries that are stable or prediction does not depend on stability
so just simply perging a stable economy against and unstable one
okay, so are we to use the other factors measure the exchange rate?
Precisely.
Actally, I can even get same data for sub-saharan Africa. But lets work on this first and see how it turns out.
great!!
I've asked Honourable Alex to join us. I believe he can make more inputs in this.
Perfect. Alexopapa
But do you have more variables to measure the exchange rate? I think just 3 is too small. All the same, I'll do something so we see. We need more data to train.
well, technically, these are major variables that econmists use. But will fish around and see how it pans out.
My approach to use::
Explore the data
well, technically, these are major variables that econmists use. But will fish around and see how it pans out.
okay, so we need to justify this in write-up with proper reference.
Remember to add the source of data too.
The write-up will be done to perfection. The source of the data is THE WORLD BANK
Will surely add it
Will do something and share!!
Yes Dr. Github is cool paa... i lik e it
I'm glad you like it. We need to introduce this to our future students.
Yeah. We shall do exactly that
Some descriptive analysis::
Time Series Projection
1. Linear Scale
2. Log Scale
Multivariate Analysis:
1. Looking at the whole datasets
Here, neither broad money nor policy rate is significant at a significant level of . The strong collinear of these predictors cause each variable to fail the t-test in the model with exception of International reserve A. Although two of the predictors fail individual t-test, overall regression is significant, meaning these 3 predictors are jointly informative, according to the F-test. Also, notice how the goodness of fit test produce an . In summary, we have observed that multicollinearity has thrown off the results of individual t-tests. On the other hand, an F-test that judges the explanatory power of all the predictors collectively, gave an improved outcome.
NB: Read multicollinearity from this link: https://en.wikipedia.org/wiki/Multicollinearity
2. Looking at Ghana only
3. Looking at South Africa only
NB:: From these, we can generate multiple regression equation.
Using an Artificial Neural Network (ANN):
Proposed Titles::
Prediction of Ghana's Exchange Rate I: A Comparative Statistical Analysis Approach
Prediction of Ghana's Exchange Rate II: A Supervised Regression Approach
Prediction of Ghana's Exchange Rate III: A Regression-based Artificial Neural network Approach
Training your model
Link to codes:: https://nbviewer.jupyter.org/urls/dl.dropbox.com/s/gw36jnqzltgz43i/exchange_rate_01.ipynb
https://nbviewer.jupyter.org/urls/dl.dropbox.com/s/vavgxjveoy9egdq/exchange_03.ipynb
Can the size of the image be made a bit larger size like (12,12)?
Change effected
On Fri, May 31, 2019 at 1:03 PM akotodanso notifications@github.com wrote:
Training your model [image: image] https://user-images.githubusercontent.com/8245353/58686698-cb83ab80-837f-11e9-87e2-66c5ca08385d.png
[image: image] https://user-images.githubusercontent.com/8245353/58686747-eeae5b00-837f-11e9-9c2f-fea757f92b9d.png
Link to codes::
https://nbviewer.jupyter.org/urls/dl.dropbox.com/s/gw36jnqzltgz43i/exchange_rate_01.ipynb
https://nbviewer.jupyter.org/urls/dl.dropbox.com/s/vavgxjveoy9egdq/exchange_03.ipynb
Can the size of the image be made a bit larger size like (12,12)?
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/philusnarh/sarkpro/issues/1?email_source=notifications&email_token=AB65A2O2ZO4HD77XFSYGDA3PYEAXRA5CNFSM4HRFLSG2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODWU5O2A#issuecomment-497669992, or mute the thread https://github.com/notifications/unsubscribe-auth/AB65A2LGLK2GNFXWE4DI3YTPYEAXRANCNFSM4HRFLSGQ .
Training your model Link to codes:: https://nbviewer.jupyter.org/urls/dl.dropbox.com/s/gw36jnqzltgz43i/exchange_rate_01.ipynb https://nbviewer.jupyter.org/urls/dl.dropbox.com/s/vavgxjveoy9egdq/exchange_03.ipynb
Can the size of the image be made a bit larger size like (12,12)?
Change effected
Left is original distribution Right: When we apply normalisation
Hello Dr. Have seen the diagram. This may suggest that prediction with the normalized data may be better than the raw data. We can do a few predictions with both and see which is better. Since the data is normalized, then we may consider using the normal distribution and see what it may give us
On Sun, Jun 2, 2019 at 12:26 PM philusnarh notifications@github.com wrote:
Left is original distribution Right: When we apply normalisation
[image: image] https://user-images.githubusercontent.com/8245353/58759909-6a9bd500-8531-11e9-9e1f-d6fcd401ad01.png
— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/philusnarh/sarkpro/issues/1?email_source=notifications&email_token=AMGSW7FWARWLORFVXRCQPODPYON5PA5CNFSM4HRFLSG2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODWXSRCI#issuecomment-498018441, or mute the thread https://github.com/notifications/unsubscribe-auth/AMGSW7DZLBNNLUMQWURJGXTPYON5PANCNFSM4HRFLSGQ .
Residual Plots
Edited Neural Network!!
https://nbviewer.jupyter.org/urls/dl.dropbox.com/s/x5r90gbqpmu2okf/exchange_03a.ipynb
Hello Guys, it has been a long time now. Hope we are good? we need to move to action soon. chooooboooiiiiii . Ghana Science Association are ready for us
At a deep neural training workshop in African Institute for Mathematical Sciences and it is good papa...... lets finish the interest rate project now and move to others
maybe we can even work on ' Marriage data and train it to see the possibility of a couple divorcing after marriage..... Interesting times ahead.
also we can work on " Time inefficiencies and its usage on productivity.....
Do we have the variables to measure and the data in particular?
On Tue, Jul 30, 2019 at 3:07 PM Sarkodie notifications@github.com wrote:
also we can work on " Time inefficiencies and its usage on productivity.....
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/philusnarh/sarkpro/issues/1?email_source=notifications&email_token=AB65A2MYAJE4XKWMIOZ3D3TQCA4K3A5CNFSM4HRFLSG2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD3D43YQ#issuecomment-516410850, or mute the thread https://github.com/notifications/unsubscribe-auth/AB65A2PZ5IO4DFCOR645W6TQCA4K3ANCNFSM4HRFLSGQ .
Not that l know of, but will ask around.
On Tue, 30 Jul 2019, 15:50 philusnarh, notifications@github.com wrote:
Do we have the variables to measure and the data in particular?
On Tue, Jul 30, 2019 at 3:07 PM Sarkodie notifications@github.com wrote:
also we can work on " Time inefficiencies and its usage on productivity.....
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub < https://github.com/philusnarh/sarkpro/issues/1?email_source=notifications&email_token=AB65A2MYAJE4XKWMIOZ3D3TQCA4K3A5CNFSM4HRFLSG2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD3D43YQ#issuecomment-516410850 , or mute the thread < https://github.com/notifications/unsubscribe-auth/AB65A2PZ5IO4DFCOR645W6TQCA4K3ANCNFSM4HRFLSGQ
.
— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/philusnarh/sarkpro/issues/1?email_source=notifications&email_token=AMGSW7F6N6LRMZINI6NNFULQCBBLLA5CNFSM4HRFLSG2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD3EA56I#issuecomment-516427513, or mute the thread https://github.com/notifications/unsubscribe-auth/AMGSW7FNUDFWXSUQSCVWOZ3QCBBLLANCNFSM4HRFLSGQ .
hmm!!, ML on marriage. Do we have the data?
On Tue, Jul 30, 2019 at 3:06 PM Sarkodie notifications@github.com wrote:
maybe we can even work on ' Marriage data and train it to see the possibility of a couple divorcing after marriage..... Interesting times ahead.
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/philusnarh/sarkpro/issues/1?email_source=notifications&email_token=AB65A2K55MO6JU6R2WIOFR3QCA4E5A5CNFSM4HRFLSG2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD3D4XFQ#issuecomment-516410262, or mute the thread https://github.com/notifications/unsubscribe-auth/AB65A2PMLRIQZUO4QLOOKQ3QCA4E5ANCNFSM4HRFLSGQ .
But for the variables, we can consider:Age difference, financial standing, religion, Academic qualification, etc etc etc
On Tue, 30 Jul 2019, 15:55 Richmond Sarkodie, rsarkodie@mucg.edu.gh wrote:
Not that l know of, but will ask around.
On Tue, 30 Jul 2019, 15:50 philusnarh, notifications@github.com wrote:
Do we have the variables to measure and the data in particular?
On Tue, Jul 30, 2019 at 3:07 PM Sarkodie notifications@github.com wrote:
also we can work on " Time inefficiencies and its usage on productivity.....
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub < https://github.com/philusnarh/sarkpro/issues/1?email_source=notifications&email_token=AB65A2MYAJE4XKWMIOZ3D3TQCA4K3A5CNFSM4HRFLSG2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD3D43YQ#issuecomment-516410850 , or mute the thread < https://github.com/notifications/unsubscribe-auth/AB65A2PZ5IO4DFCOR645W6TQCA4K3ANCNFSM4HRFLSGQ
.
— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/philusnarh/sarkpro/issues/1?email_source=notifications&email_token=AMGSW7F6N6LRMZINI6NNFULQCBBLLA5CNFSM4HRFLSG2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD3EA56I#issuecomment-516427513, or mute the thread https://github.com/notifications/unsubscribe-auth/AMGSW7FNUDFWXSUQSCVWOZ3QCBBLLANCNFSM4HRFLSGQ .
Great, when is their next submission? For now, the pressure they small for here. I have 3 months to return home. We can put in these things in the month of December.
On Tue, Jul 30, 2019 at 2:57 PM Sarkodie notifications@github.com wrote:
Hello Guys, it has been a long time now. Hope we are good? we need to move to action soon. chooooboooiiiiii . Ghana Science Association are ready for us
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/philusnarh/sarkpro/issues/1?email_source=notifications&email_token=AB65A2NPP5M24TB2P37PYR3QCA3FDA5CNFSM4HRFLSG2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD3D4AHQ#issuecomment-516407326, or mute the thread https://github.com/notifications/unsubscribe-auth/AB65A2M2VUPDSGUO3TO7KRDQCA3FDANCNFSM4HRFLSGQ .
If data is not readily available, we can collect it ourselves
On Tue, 30 Jul 2019, 15:53 oseisarkodie, notifications@github.com wrote:
But for the variables, we can consider:Age difference, financial standing, religion, Academic qualification, etc etc etc
On Tue, 30 Jul 2019, 15:55 Richmond Sarkodie, rsarkodie@mucg.edu.gh wrote:
Not that l know of, but will ask around.
On Tue, 30 Jul 2019, 15:50 philusnarh, notifications@github.com wrote:
Do we have the variables to measure and the data in particular?
On Tue, Jul 30, 2019 at 3:07 PM Sarkodie notifications@github.com wrote:
also we can work on " Time inefficiencies and its usage on productivity.....
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <
, or mute the thread <
https://github.com/notifications/unsubscribe-auth/AB65A2PZ5IO4DFCOR645W6TQCA4K3ANCNFSM4HRFLSGQ
.
— You are receiving this because you commented. Reply to this email directly, view it on GitHub < https://github.com/philusnarh/sarkpro/issues/1?email_source=notifications&email_token=AMGSW7F6N6LRMZINI6NNFULQCBBLLA5CNFSM4HRFLSG2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD3EA56I#issuecomment-516427513 , or mute the thread < https://github.com/notifications/unsubscribe-auth/AMGSW7FNUDFWXSUQSCVWOZ3QCBBLLANCNFSM4HRFLSGQ
.
— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/philusnarh/sarkpro/issues/1?email_source=notifications&email_token=AF74E4OQM62WTP5MPQLAFYDQCBBWBA5CNFSM4HRFLSG2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD3EBHQY#issuecomment-516428739, or mute the thread https://github.com/notifications/unsubscribe-auth/AF74E4IABK62MNOKSJDISFLQCBBWBANCNFSM4HRFLSGQ .
For now they are working on their online submission system which will be out soon. With that one can publish anytime their work is accepted. Then later they will compile into a book
On Tue, 30 Jul 2019, 15:56 philusnarh, notifications@github.com wrote:
Great, when is their next submission? For now, the pressure they small for here. I have 3 months to return home. We can put in these things in the month of December.
On Tue, Jul 30, 2019 at 2:57 PM Sarkodie notifications@github.com wrote:
Hello Guys, it has been a long time now. Hope we are good? we need to move to action soon. chooooboooiiiiii . Ghana Science Association are ready for us
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub < https://github.com/philusnarh/sarkpro/issues/1?email_source=notifications&email_token=AB65A2NPP5M24TB2P37PYR3QCA3FDA5CNFSM4HRFLSG2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD3D4AHQ#issuecomment-516407326 , or mute the thread < https://github.com/notifications/unsubscribe-auth/AB65A2M2VUPDSGUO3TO7KRDQCA3FDANCNFSM4HRFLSGQ
.
— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/philusnarh/sarkpro/issues/1?email_source=notifications&email_token=AMGSW7AMJFL655GQUMCKIWDQCBCBDA5CNFSM4HRFLSG2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD3EBSKY#issuecomment-516430123, or mute the thread https://github.com/notifications/unsubscribe-auth/AMGSW7ABVGL5QR4HZTNRO6LQCBCBDANCNFSM4HRFLSGQ .
But for the variables, we can consider:Age difference, financial standing, religion, Academic qualification, etc etc etc … On Tue, 30 Jul 2019, 15:55 Richmond Sarkodie, @.> wrote: Not that l know of, but will ask around. On Tue, 30 Jul 2019, 15:50 philusnarh, @.> wrote: > Do we have the variables to measure and the data in particular? > > On Tue, Jul 30, 2019 at 3:07 PM Sarkodie @.***> > wrote: > > > also we can work on " Time inefficiencies and its usage on > > productivity..... > > > > — > > You are receiving this because you authored the thread. > > Reply to this email directly, view it on GitHub > > < > #1?email_source=notifications&email_token=AB65A2MYAJE4XKWMIOZ3D3TQCA4K3A5CNFSM4HRFLSG2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD3D43YQ#issuecomment-516410850 > >, > > or mute the thread > > < > https://github.com/notifications/unsubscribe-auth/AB65A2PZ5IO4DFCOR645W6TQCA4K3ANCNFSM4HRFLSGQ > > > > . > > > > — > You are receiving this because you commented. > Reply to this email directly, view it on GitHub > <#1?email_source=notifications&email_token=AMGSW7F6N6LRMZINI6NNFULQCBBLLA5CNFSM4HRFLSG2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD3EA56I#issuecomment-516427513>, > or mute the thread > https://github.com/notifications/unsubscribe-auth/AMGSW7FNUDFWXSUQSCVWOZ3QCBBLLANCNFSM4HRFLSGQ > . >
So from primary source!!. We need like 1000plus data size oooh. This will be expensive. Unless we use the online approach by developing a google form from a standard questionnaire!!!
I think we should also look at projects that probably will give us some funds.
eg.
okay, that's an improvement. Thanks for the info.
On Tue, Jul 30, 2019 at 3:59 PM oseisarkodie notifications@github.com wrote:
For now they are working on their online submission system which will be out soon. With that one can publish anytime their work is accepted. Then later they will compile into a book
On Tue, 30 Jul 2019, 15:56 philusnarh, notifications@github.com wrote:
Great, when is their next submission? For now, the pressure they small for here. I have 3 months to return home. We can put in these things in the month of December.
On Tue, Jul 30, 2019 at 2:57 PM Sarkodie notifications@github.com wrote:
Hello Guys, it has been a long time now. Hope we are good? we need to move to action soon. chooooboooiiiiii . Ghana Science Association are ready for us
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <
, or mute the thread <
https://github.com/notifications/unsubscribe-auth/AB65A2M2VUPDSGUO3TO7KRDQCA3FDANCNFSM4HRFLSGQ
.
— You are receiving this because you commented. Reply to this email directly, view it on GitHub < https://github.com/philusnarh/sarkpro/issues/1?email_source=notifications&email_token=AMGSW7AMJFL655GQUMCKIWDQCBCBDA5CNFSM4HRFLSG2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD3EBSKY#issuecomment-516430123 , or mute the thread < https://github.com/notifications/unsubscribe-auth/AMGSW7ABVGL5QR4HZTNRO6LQCBCBDANCNFSM4HRFLSGQ
.
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/philusnarh/sarkpro/issues/1?email_source=notifications&email_token=AB65A2LIQ3TJAZ2TFSA5VMTQCBCKJA5CNFSM4HRFLSG2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD3EB2ZQ#issuecomment-516431206, or mute the thread https://github.com/notifications/unsubscribe-auth/AB65A2MZWRVDIWDAPJKKD73QCBCKJANCNFSM4HRFLSGQ .
Right. Man needs cash
On Tue, 30 Jul 2019, 16:08 philusnarh, notifications@github.com wrote:
okay, that's an improvement. Thanks for the info.
On Tue, Jul 30, 2019 at 3:59 PM oseisarkodie notifications@github.com wrote:
For now they are working on their online submission system which will be out soon. With that one can publish anytime their work is accepted. Then later they will compile into a book
On Tue, 30 Jul 2019, 15:56 philusnarh, notifications@github.com wrote:
Great, when is their next submission? For now, the pressure they small for here. I have 3 months to return home. We can put in these things in the month of December.
On Tue, Jul 30, 2019 at 2:57 PM Sarkodie notifications@github.com wrote:
Hello Guys, it has been a long time now. Hope we are good? we need to move to action soon. chooooboooiiiiii . Ghana Science Association are ready for us
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <
, or mute the thread <
https://github.com/notifications/unsubscribe-auth/AB65A2M2VUPDSGUO3TO7KRDQCA3FDANCNFSM4HRFLSGQ
.
— You are receiving this because you commented. Reply to this email directly, view it on GitHub <
, or mute the thread <
https://github.com/notifications/unsubscribe-auth/AMGSW7ABVGL5QR4HZTNRO6LQCBCBDANCNFSM4HRFLSGQ
.
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub < https://github.com/philusnarh/sarkpro/issues/1?email_source=notifications&email_token=AB65A2LIQ3TJAZ2TFSA5VMTQCBCKJA5CNFSM4HRFLSG2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD3EB2ZQ#issuecomment-516431206 , or mute the thread < https://github.com/notifications/unsubscribe-auth/AB65A2MZWRVDIWDAPJKKD73QCBCKJANCNFSM4HRFLSGQ
.
— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/philusnarh/sarkpro/issues/1?email_source=notifications&email_token=AMGSW7D7K6MA3RAFHB2P4KTQCBDNRA5CNFSM4HRFLSG2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD3ECZ6I#issuecomment-516435193, or mute the thread https://github.com/notifications/unsubscribe-auth/AMGSW7CKYN2T7FHMHL5K4G3QCBDNRANCNFSM4HRFLSGQ .
Masa, make steady oo. You want make Ghanaian women beat we up?
On Tue, 30 Jul 2019, 15:06 Sarkodie, notifications@github.com wrote:
maybe we can even work on ' Marriage data and train it to see the possibility of a couple divorcing after marriage..... Interesting times ahead.
— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/philusnarh/sarkpro/issues/1?email_source=notifications&email_token=ACT3S4GJ5YV5UBLGHRKXAFLQCA4E5A5CNFSM4HRFLSG2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD3D4XFQ#issuecomment-516410262, or mute the thread https://github.com/notifications/unsubscribe-auth/ACT3S4FYXY7RLXHFQVQO4H3QCA4E5ANCNFSM4HRFLSGQ .
Okay, programming they inside or pure theoretical.
On Tue, 30 Jul 2019, 2:59 PM Sarkodie notifications@github.com wrote:
At a deep neural training workshop in African Institute for Mathematical Sciences and it is good papa...... lets finish the interest rate project now and move to others
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/philusnarh/sarkpro/issues/1?email_source=notifications&email_token=AB65A2JBQBG22RCBYWJGED3QCA3LBA5CNFSM4HRFLSG2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD3D4ERY#issuecomment-516407879, or mute the thread https://github.com/notifications/unsubscribe-auth/AB65A2N6ZYTYH6TK6NAUFZLQCA3LBANCNFSM4HRFLSGQ .
So kindly list the project outline of this research