Skip to content

A
alpha-mind
Commit 1f57b3df authored by Dr.李's avatar Dr.李
Browse files
Options

added example

parent 04f81b19
Show whitespace changes
Inline Side-by-side
Showing with 384 additions and 0 deletions
notebooks/Example 6 - Target Volatility Builder.ipynb 0 → 100644
View file @ 1f57b3df
{
"cells": [
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"%matplotlib inline\n",
"import numpy as np\n",
"import pandas as pd\n",
"from matplotlib import pyplot as plt\n",
"from PyFin.api import *\n",
"from alphamind.api import *\n",
"from alphamind.portfolio.meanvariancebuilder import target_vol_builder\n",
"\n",
"plt.style.use('ggplot')"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## 1. Single Day Analysis\n",
"-----------------------"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"ref_date = '2018-02-08'\n",
"engine = SqlEngine('postgres+psycopg2://postgres:we083826@192.168.0.102/alpha')\n",
"universe = Universe('custom', ['zz800'])"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"codes = engine.fetch_codes(ref_date, universe)\n",
"total_data = engine.fetch_data(ref_date, 'ep_q', codes, 905, industry='sw_adj')\n",
"all_styles = risk_styles + industry_styles + ['COUNTRY']"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"risk_cov = total_data['risk_cov'][all_styles].values\n",
"factor = total_data['factor']\n",
"risk_exposure = factor[all_styles].values\n",
"special_risk = factor['s_srisk'].values"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"sec_cov = risk_exposure @ risk_cov @ risk_exposure.T / 10000 + np.diag(special_risk ** 2) / 10000\n",
"sec_cov_df = pd.DataFrame(sec_cov, index=codes, columns=codes)\n",
"sec_cov_df.iloc[:5, :5]"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Portfolio Construction\n",
"\n",
"* using `ep_q` factor as alpha factor;\n",
"* short selling is forbiden;\n",
"* target of volatility for the activate weight is setting at 5% annually level."
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"er = factor['ep_q'].values\n",
"bm = factor['weight'].values\n",
"lbound = np.zeros(len(er))\n",
"ubound = bm + 0.05\n",
"cons_mat = np.ones((len(er), 1))\n",
"risk_targets = (bm.sum(), bm.sum())\n",
"target_vol = 0.045\n",
"\n",
"status, p_er, p_weight = \\\n",
" target_vol_builder(er, sec_cov, bm, lbound, ubound, cons_mat, risk_targets, target_vol, target_vol)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"# check the result\n",
"print(f\"total weight is {p_weight.sum(): .4f}\")\n",
"print(f\"portfolio activate weight forecasting vol is {np.sqrt((p_weight - bm) @ sec_cov @ (p_weight - bm)):.2f}\")\n",
"print(f\"portfolio expected return is {p_weight @ er:.4f} comparing with benchmark er {bm @ er:.4f}\")"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## 2. Porfolio Construction: 2010 ~ 2018\n",
"-------------------------------"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"\"\"\"\n",
"Back test parameter settings\n",
"\"\"\"\n",
"\n",
"start_date = '2011-01-01'\n",
"end_date = '2018-02-14'\n",
"\n",
"freq = '10b'\n",
"industry_lower = 1.0\n",
"industry_upper = 1.0\n",
"neutralized_risk = ['SIZE'] + industry_styles\n",
"industry_name = 'sw_adj'\n",
"industry_level = 1\n",
"batch = 0\n",
"horizon = map_freq(freq)\n",
"universe = Universe(\"custom\", ['zz500'])\n",
"data_source = 'postgres+psycopg2://postgres:we083826@192.168.0.102/alpha'\n",
"benchmark_code = 905\n",
"target_vol = 0.05\n",
"\n",
"executor = NaiveExecutor()\n",
"ref_dates = makeSchedule(start_date, end_date, freq, 'china.sse')\n",
"engine = SqlEngine(data_source)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"\"\"\"\n",
"Factor Model\n",
"\"\"\"\n",
"\n",
"alpha_factors = {'f01': LAST('ep_q')}\n",
"\n",
"weights = dict(f01=1.)\n",
"\n",
"alpha_model = ConstLinearModel(features=alpha_factors, weights=weights)\n",
"\n",
"def predict_worker(params):\n",
" data_meta = DataMeta(freq=freq,\n",
" universe=universe,\n",
" batch=batch,\n",
" neutralized_risk=neutralized_risk,\n",
" risk_model='short',\n",
" pre_process=[winsorize_normal, standardize],\n",
" post_process=[winsorize_normal, standardize],\n",
" warm_start=0,\n",
" data_source=data_source)\n",
" ref_date, model = params\n",
" er = predict_by_model(ref_date, model, data_meta)\n",
" return er"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"%%time\n",
"\n",
"\"\"\"\n",
"Predicting Phase\n",
"\"\"\"\n",
"\n",
"from dask.distributed import Client\n",
"client = Client('192.168.0.102:8786')\n",
"\n",
"tasks = client.map(predict_worker, [(d.strftime('%Y-%m-%d'), alpha_model) for d in ref_dates], pure=False)\n",
"predicts = client.gather(tasks)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"\"\"\"\n",
"Shared Data\n",
"\"\"\"\n",
"\n",
"constraint_risk = ['SIZE', 'SIZENL', 'BETA']\n",
"total_risk_names = constraint_risk + ['total']\n",
"\n",
"b_type = []\n",
"l_val = []\n",
"u_val = []\n",
"\n",
"for name in total_risk_names:\n",
" b_type.append(BoundaryType.ABSOLUTE)\n",
" l_val.append(0.0)\n",
" u_val.append(0.0)\n",
" \n",
"bounds = create_box_bounds(total_risk_names, b_type, l_val, u_val)\n",
"industry_total = engine.fetch_industry_matrix_range(universe, dates=ref_dates, category=industry_name, level=industry_level)\n",
"benchmark_total = engine.fetch_benchmark_range(dates=ref_dates, benchmark=benchmark_code)\n",
"risk_cov_total, risk_exposure_total = engine.fetch_risk_model_range(universe, dates=ref_dates)\n",
"index_return = engine.fetch_dx_return_index_range(benchmark_code, start_date, end_date, horizon=horizon, offset=1).set_index('trade_date')"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"# rebalance\n",
"\n",
"def create_scenario(target_vol):\n",
" \n",
" all_styles = risk_styles + industry_styles + macro_styles\n",
" previous_pos = pd.DataFrame()\n",
" rets = []\n",
" turn_overs = []\n",
" leverags = []\n",
" ics = []\n",
"\n",
" for i, ref_date in enumerate(ref_dates):\n",
" ref_date = ref_date.strftime('%Y-%m-%d')\n",
" industry_matrix = industry_total[industry_total.trade_date == ref_date]\n",
" benchmark_w = benchmark_total[benchmark_total.trade_date == ref_date]\n",
" risk_exposure = risk_exposure_total[risk_exposure_total.trade_date == ref_date]\n",
" risk_cov = risk_cov_total[risk_cov_total.trade_date == ref_date]\n",
" \n",
" total_data = pd.merge(industry_matrix, benchmark_w, on=['code'], how='left').fillna(0.)\n",
" total_data = pd.merge(total_data, risk_exposure, on=['code'])\n",
" total_data = total_data.dropna()\n",
" codes = total_data.code.values.tolist()\n",
" \n",
" risk_exposure = total_data[all_styles].values\n",
" risk_cov = risk_cov[all_styles].values\n",
" special_risk = total_data['s_srisk'].values\n",
" sec_cov = risk_exposure @ risk_cov @ risk_exposure.T / 10000 + np.diag(special_risk ** 2) / 10000\n",
"\n",
" benchmark_w = total_data.weight.values\n",
" \n",
" total_risk_exp = np.concatenate([total_data[constraint_risk].values.astype(float),\n",
" np.ones((len(benchmark_w),1))],\n",
" axis=1)\n",
" total_risk_exp = pd.DataFrame(total_risk_exp, columns=total_risk_names)\n",
" constraints = LinearConstraints(bounds, total_risk_exp, benchmark_w)\n",
" \n",
" lbound = np.zeros(len(total_data))\n",
" ubound = 0.05 + benchmark_w\n",
"\n",
" er = predicts[i].loc[codes].values.flatten()\n",
" cons_mat = np.ones((len(er), 1))\n",
" risk_target = (benchmark_w.sum(), benchmark_w.sum())\n",
" \n",
" try:\n",
" target_pos, _ = er_portfolio_analysis(er,\n",
" industry_matrix.industry_name.values,\n",
" None,\n",
" constraints,\n",
" False,\n",
" benchmark_w,\n",
" method='tv',\n",
" lbound=lbound,\n",
" ubound=ubound,\n",
" cov=sec_cov,\n",
" target_vol=target_vol)\n",
" except:\n",
" import pdb\n",
" pdb.set_trace()\n",
"\n",
" target_pos['code'] = codes\n",
" turn_over, executed_pos = executor.execute(target_pos=target_pos)\n",
"\n",
" executed_codes = executed_pos.code.tolist()\n",
" dx_returns = engine.fetch_dx_return(ref_date, executed_codes, horizon=horizon, offset=1)\n",
" result = pd.merge(executed_pos, total_data[['code', 'weight']], on=['code'], how='inner')\n",
" result = pd.merge(result, dx_returns, on=['code'])\n",
" \n",
" excess_return = np.exp(result.dx.values) - 1. - index_return.loc[ref_date, 'dx']\n",
" raw_weight = result.weight_x.values\n",
" activate_weight = raw_weight - result.weight_y.values\n",
" ret = raw_weight @ excess_return\n",
" risk_adjusted_ic = np.corrcoef(excess_return, activate_weight)[0, 1]\n",
" rets.append(np.log(1. + ret))\n",
" ics.append(risk_adjusted_ic)\n",
" executor.set_current(executed_pos)\n",
" turn_overs.append(turn_over)\n",
" \n",
" leverage = raw_weight.sum()\n",
" leverags.append(leverage)\n",
" alpha_logger.info(f\"{ref_date} is finished with expected vol {np.sqrt((target_pos.weight.values - benchmark_w) @ sec_cov @ (target_pos.weight.values - benchmark_w)):.2f}\")\n",
"\n",
" ret_df = pd.DataFrame({'returns': rets, 'turn_over': turn_overs, 'IC': ics, 'leverage': leverags}, index=ref_dates)\n",
"\n",
" ret_df.loc[advanceDateByCalendar('china.sse', ref_dates[-1], freq)] = 0.\n",
" ret_df = ret_df.shift(1)\n",
" ret_df.iloc[0] = 0.\n",
" ret_df['tc_cost'] = ret_df.turn_over * 0.002\n",
" return ret_df"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"ret_df = create_scenario(target_vol)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"ret_df[['returns', 'tc_cost']].cumsum().plot(figsize=(12, 6),\n",
" title='Fixed freq rebalanced: {0} with benchmark {1}'.format(freq, 905),\n",
" secondary_y='tc_cost')"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.6.4"
}
},
"nbformat": 4,
"nbformat_minor": 2
}
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment