delta-barth/delta-barth-py
1
0
Fork 0
Code Issues 3 Pull Requests Actions Packages Projects 3 Releases 3 Wiki Activity

src/delta_barth/analysis/forecast.py aktualisiert

Browse Source
This commit is contained in:
frasu 2025-04-13 14:45:55 +00:00
parent 8501f551b2
commit 0eb39deec5
1 changed files with 17 additions and 16 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

33
src/delta_barth/analysis/forecast.py
View File

@ -210,7 +210,18 @@ def _process_sales(
df_cust["jahr"] = df_cust[DATE_FEAT].dt.year df_cust["jahr"] = df_cust[DATE_FEAT].dt.year
df_cust["monat"] = df_cust[DATE_FEAT].dt.month df_cust["monat"] = df_cust[DATE_FEAT].dt.month
monthly_sum = df_cust.groupby(["jahr", "monat"])[SALES_FEAT].sum().reset_index() current_year = datetime.now().year
current_month = datetime.now().month
years = range(df_cust["jahr"].min(), current_year + 1)
old_monthly_sum = df_cust.groupby(["jahr", "monat"])[SALES_FEAT].sum().reset_index()
all_month_year_combinations = pd.DataFrame(
[(year, month) for year in years for month in range(1, 13) if (year < current_year or (year == current_year and month <= current_month))], columns=["jahr", "monat"]
)
monthly_sum = pd.merge(all_month_year_combinations, old_monthly_sum, on=["jahr", "monat"], how="left")
monthly_sum[SALES_FEAT] = monthly_sum[SALES_FEAT].fillna(0)
monthly_sum[DATE_FEAT] = ( monthly_sum[DATE_FEAT] = (
monthly_sum["monat"].astype(str) + "." + monthly_sum["jahr"].astype(str) monthly_sum["monat"].astype(str) + "." + monthly_sum["jahr"].astype(str)
) )
@ -220,7 +231,6 @@ def _process_sales(
features = ["jahr", "monat"] features = ["jahr", "monat"]
target = SALES_FEAT target = SALES_FEAT
# ?? --- new: dates and forecast
last_date = pd.to_datetime(datetime.datetime.now().strftime("%m.%Y"), format="%m.%Y") last_date = pd.to_datetime(datetime.datetime.now().strftime("%m.%Y"), format="%m.%Y")
future_dates = pd.date_range( future_dates = pd.date_range(
start=last_date + pd.DateOffset(months=1), periods=6, freq="MS" start=last_date + pd.DateOffset(months=1), periods=6, freq="MS"
@ -230,7 +240,7 @@ def _process_sales(
# Randomized Search # Randomized Search
kfold = KFold(n_splits=5, shuffle=True) kfold = KFold(n_splits=5, shuffle=True)
params: ParamSearchXGBRegressor = { params: ParamSearchXGBRegressor = {
"n_estimators": scipy.stats.poisson(mu=1000), "n_estimators": scipy.stats.poisson(mu=100),
"learning_rate": [0.03, 0.04, 0.05], "learning_rate": [0.03, 0.04, 0.05],
"max_depth": range(2, 9), "max_depth": range(2, 9),
"min_child_weight": range(1, 5), "min_child_weight": range(1, 5),
@ -240,27 +250,19 @@ def _process_sales(
"early_stopping_rounds": [20, 50], "early_stopping_rounds": [20, 50],
} }
# ?? --- new: best_estimator (internal usage only)
best_estimator = None best_estimator = None
best_params: BestParametersXGBRegressor | None = None best_params: BestParametersXGBRegressor | None = None
best_score_mae: float | None = float("inf") best_score_mae: float | None = float("inf")
best_score_r2: float | None = None best_score_r2: float | None = None
best_start_year: int | None = None best_start_year: int | None = None
too_few_month_points: bool = True too_few_month_points: bool = True
# forecast: pd.DataFrame | None = None
# TODO: write routine to pad missing values in datetime row
# TODO problem: continuous timeline expected, but values can be empty for multiple months
# TODO: therefore, stepping with fixed value n does not result in timedelta of n episodes
# Option A: pad data frame with zero values --> could impede forecast algorithm
# Option B: calculate next index based on timedelta
stride = dopt_basics.datetime.timedelta_from_val(365, TimeUnitsTimedelta.DAYS) stride = dopt_basics.datetime.timedelta_from_val(365, TimeUnitsTimedelta.DAYS)
dates = cast(pd.DatetimeIndex, monthly_sum.index) dates = cast(pd.DatetimeIndex, monthly_sum.index)
min_date = dates.min() min_date = dates.min()
# ?? --- new: use monthly basis for time windows
# baseline: 3 years - 36 months # baseline: 3 years - 36 months
starting_date = datetime.datetime.now() - relativedelta(months=36) starting_date = datetime.datetime.now() - relativedelta(months=36)
# starting_date = dates.max() - relativedelta(months=36)
def get_index_date( def get_index_date(
dates: pd.DatetimeIndex, dates: pd.DatetimeIndex,
@ -307,10 +309,9 @@ def _process_sales(
X_train, X_test = train[features], test[features] X_train, X_test = train[features], test[features]
y_train, y_test = train[target], test[target] y_train, y_test = train[target], test[target]
# ?? --- new: adapted condition to fit new for-loop
# test set size fixed at 6 --> first iteration: baseline - 6 entries # test set size fixed at 6 --> first iteration: baseline - 6 entries
# for each new year 10 new data points needed # for each new year 10 new data points (i.e., sales strictly positive) needed
if len(train) >= 30 + 10 * step: if len(train[train[SALES_FEAT] > 0]) >= 30 + 10 * step:
too_few_month_points = False too_few_month_points = False
rand = RandomizedSearchCV( rand = RandomizedSearchCV(
@ -333,7 +334,7 @@ def _process_sales(
best_params = cast(BestParametersXGBRegressor, rand.best_params_) best_params = cast(BestParametersXGBRegressor, rand.best_params_)
best_score_mae = error best_score_mae = error
best_score_r2 = cast(float, r2_score(y_test, y_pred)) best_score_r2 = cast(float, r2_score(y_test, y_pred))
# --- new: use first_date for best_start_year # --- new: use target_date for best_start_year
best_start_year = target_date.year best_start_year = target_date.year
# --- new: store best_estimator # --- new: store best_estimator
best_estimator = copy.copy(rand.best_estimator_) best_estimator = copy.copy(rand.best_estimator_)