Eu resolvi melhorar a previsão e o modelo pra tal sobre o Linux no desktop, uma realidade inegável. Então escrevi algo em python pra fazer isso pra mim o trabalho e usar de algum modelo já pronto.
(venv) helio@goosfraba ~/D/linux-desktop-dominance-forecast (main)> ./forecasting.py
Windows OS X Unknown Linux Chrome OS iOS Android \
Date
2009-01-01 95.42 3.68 0.17 0.64 0.0 0.0 0.0
2009-02-01 95.39 3.76 0.14 0.62 0.0 0.0 0.0
2009-03-01 95.22 3.87 0.16 0.65 0.0 0.0 0.0
2009-04-01 95.13 3.92 0.17 0.66 0.0 0.0 0.0
2009-05-01 95.25 3.75 0.24 0.65 0.0 0.0 0.0
Playstation Other
Date
2009-01-01 0.08 0.02
2009-02-01 0.07 0.02
2009-03-01 0.08 0.02
2009-04-01 0.10 0.02
2009-05-01 0.09 0.02
Windows OS X Unknown Linux Chrome OS iOS Android \
Date
2024-01-01 73.00 16.11 5.33 3.77 1.78 0.0 0.0
2024-02-01 72.17 15.42 6.10 4.03 2.27 0.0 0.0
2024-03-01 72.47 14.68 6.52 4.05 2.27 0.0 0.0
2024-04-01 73.50 14.70 5.34 3.88 2.56 0.0 0.0
2024-05-01 73.91 14.90 4.87 3.77 2.54 0.0 0.0
2024-06-01 72.81 14.97 6.23 4.05 1.93 0.0 0.0
2024-07-01 72.10 14.92 7.13 4.44 1.41 0.0 0.0
Playstation Other
Date
2024-01-01 0.0 0.01
2024-02-01 0.0 0.01
2024-03-01 0.0 0.01
2024-04-01 0.0 0.01
2024-05-01 0.0 0.01
2024-06-01 0.0 0.01
2024-07-01 0.0 0.01
/home/helio/DEVEL/linux-desktop-dominance-forecast/venv/lib/python3.12/site-packages/statsmodels/tsa/base/tsa_model.py:473: ValueWarning: No frequency information was provided, so inferred frequency MS will be used.
self._init_dates(dates, freq)
/home/helio/DEVEL/linux-desktop-dominance-forecast/venv/lib/python3.12/site-packages/statsmodels/tsa/base/tsa_model.py:473: ValueWarning: No frequency information was provided, so inferred frequency MS will be used.
self._init_dates(dates, freq)
/home/helio/DEVEL/linux-desktop-dominance-forecast/venv/lib/python3.12/site-packages/statsmodels/tsa/statespace/sarimax.py:978: UserWarning: Non-invertible starting MA parameters found. Using zeros as starting parameters.
warn('Non-invertible starting MA parameters found.'
RUNNING THE L-BFGS-B CODE
* * *
Machine precision = 2.220D-16
N = 11 M = 10
This problem is unconstrained.
At X0 0 variables are exactly at the bounds
At iterate 0 f= 2.00575D-01 |proj g|= 1.24633D+00
At iterate 5 f= -1.49635D-01 |proj g|= 1.77236D-01
At iterate 10 f= -3.64537D-01 |proj g|= 1.24837D-01
At iterate 15 f= -3.84843D-01 |proj g|= 1.22570D-01
At iterate 20 f= -4.20877D-01 |proj g|= 9.71167D-02
At iterate 25 f= -4.29351D-01 |proj g|= 1.13565D-01
At iterate 30 f= -4.33425D-01 |proj g|= 9.06436D-02
At iterate 35 f= -4.34142D-01 |proj g|= 3.98871D-02
At iterate 40 f= -4.36192D-01 |proj g|= 4.26757D-01
At iterate 45 f= -4.37801D-01 |proj g|= 6.82859D-02
At iterate 50 f= -4.37938D-01 |proj g|= 1.67285D-02
* * *
Tit = total number of iterations
Tnf = total number of function evaluations
Tnint = total number of segments explored during Cauchy searches
Skip = number of BFGS updates skipped
Nact = number of active bounds at final generalized Cauchy point
Projg = norm of the final projected gradient
F = final function value
* * *
N Tit Tnf Tnint Skip Nact Projg F
11 50 55 1 0 0 1.673D-02 -4.379D-01
F = -0.43793754789434586
STOP: TOTAL NO. of ITERATIONS REACHED LIMIT
/home/helio/DEVEL/linux-desktop-dominance-forecast/venv/lib/python3.12/site-packages/statsmodels/base/model.py:607:
ConvergenceWarning: Maximum Likelihood optimization failed to converge. Check mle_retvals
warnings.warn("Maximum Likelihood optimization failed to "
The year of Linux on the Desktop: 2036-11-30 00:00:00
Se ignorarmos os pequenos erros e avisos que aparecem, coisa pouca e irrelevante como valor divergir demais, podemos ver claramente quando o ano do Linux no desktop acontece: 2036-11-30
Então a previsão revisada é que logo estaremos em todos os lugares. Aguardem-nos!
E claro que publiquei isso tudo no GitHub:
E usei o seguinte artigo como referência (e código, diga-se de passagem):
Bom ano do Linux no desktop pra todos vocês.
