Skip to content

A
alpha-mind
Commit e9a25f85 authored by Dr.李's avatar Dr.李
Browse files
Options

update example

parent a9579e26
Hide whitespace changes
Inline Side-by-side
Showing with 165 additions and 34 deletions
notebooks/Quick Start 4 - Step By Step Alpha Modeling.ipynb
View file @ e9a25f85
......@@ -33,9 +33,25 @@
"rc('legend', **{'frameon': False})"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## 1. 组合构造要求\n",
"--------------------------\n",
"\n",
"* 股票池：中证800\n",
"* 交易日：2019年1月15日\n",
"* alpha因子：EPS\n",
"* 行业分类：申万行业一级分类\n",
"* 基准指数：中证800指数\n",
"* 行业暴露：各行业保持中性\n",
"* 个股暴露：个股最大主动暴露不超过2%"
]
},
{
"cell_type": "code",
"execution_count": 95,
"execution_count": 140,
"metadata": {},
"outputs": [],
"source": [
......@@ -49,16 +65,32 @@
"con = sa.create_engine(os.environ['DB_URI'])"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## 2. 获取基本数据\n",
"----------------------\n",
"\n",
"获取在2019年1月15日中证800成分股的EPS因子值"
]
},
{
"cell_type": "code",
"execution_count": 96,
"execution_count": 142,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Wall time: 457 ms\n"
" code EPS\n",
"0 1 1.5885\n",
"1 2 1.6891\n",
"2 6 0.4970\n",
"3 8 0.0592\n",
"4 9 0.0612\n",
"Wall time: 150 ms\n"
]
}
],
......@@ -73,19 +105,32 @@
" getattr(Universe, universe_name) == 1))\n",
"\n",
"factors = pd.read_sql(query, con=con)\n",
"factors.head()"
"print(factors.head())"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"获取在2019年1月15日中证800成分股的行业分类"
]
},
{
"cell_type": "code",
"execution_count": 97,
"execution_count": 143,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Wall time: 160 ms\n"
" code industryName1\n",
"0 1 银行\n",
"1 2 房地产\n",
"2 6 房地产\n",
"3 8 机械设备\n",
"4 9 综合\n",
"Wall time: 140 ms\n"
]
}
],
......@@ -101,72 +146,123 @@
" getattr(Universe, universe_name) == 1))\n",
"\n",
"industry = pd.read_sql(query, con=con)\n",
"industry.head()"
"print(industry.head())"
]
},
{
"cell_type": "code",
"execution_count": 98,
"cell_type": "markdown",
"metadata": {},
"outputs": [],
"source": [
"df = pd.merge(factors, industry, on=['code'], how='inner').dropna()\n",
"industry_dummy = pd.get_dummies(df.industryName1)"
"获取在2019年1月15日中证800成分股的指数权重"
]
},
{
"cell_type": "code",
"execution_count": 99,
"execution_count": 148,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Wall time: 1 ms\n"
"Wall time: 413 ms\n"
]
}
],
"source": [
"%%time\n",
"\n",
"df['neutralized_factor'] = neutralize(industry_dummy.values.astype(float), df[factor].values).flatten()"
"big_table = outerjoin(IndexComponent, Universe, and_(IndexComponent.trade_date == Universe.trade_date,\n",
" IndexComponent.code == Universe.code))\n",
"\n",
"query = select([IndexComponent.code, (IndexComponent.weight / 100.).label('index_weight')]) \\\n",
" .where(and_(IndexComponent.trade_date == trade_date,\n",
" IndexComponent.indexCode == benchmark_code))\n",
"\n",
"index_components = pd.read_sql(query, con=con)\n",
"print(index_components.head())"
]
},
{
"cell_type": "code",
"execution_count": 149,
"metadata": {},
"outputs": [],
"source": [
"df = pd.merge(factors, industry, on=['code'], how='inner').dropna()\n",
"df = pd.merge(df, index_components, on=['code'], how='inner').dropna()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## 3. 因子中性化\n",
"-------------------"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"将行业数据（categorical）数据转为dummy矩阵"
]
},
{
"cell_type": "code",
"execution_count": 100,
"execution_count": 98,
"metadata": {},
"outputs": [],
"source": [
"industry_dummy = pd.get_dummies(df.industryName1)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"使用行业dummy矩阵对因子做中性化，得到**行业中性化后**的因子：`neutralized_factor`"
]
},
{
"cell_type": "code",
"execution_count": 152,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Wall time: 113 ms\n"
" code neutralized_factor\n",
"0 1 0.253364\n",
"1 2 0.990502\n",
"2 6 -0.201598\n",
"3 8 -0.396290\n",
"4 9 -0.168950\n",
"Wall time: 2.96 ms\n"
]
}
],
"source": [
"%%time\n",
"\n",
"big_table = outerjoin(IndexComponent, Universe, and_(IndexComponent.trade_date == Universe.trade_date,\n",
" IndexComponent.code == Universe.code))\n",
"\n",
"query = select([IndexComponent.code, (IndexComponent.weight / 100.).label('index_weight')]) \\\n",
" .where(and_(IndexComponent.trade_date == trade_date,\n",
" IndexComponent.indexCode == benchmark_code))\n",
"\n",
"index_components = pd.read_sql(query, con=con)\n",
"index_components.head()"
"df['neutralized_factor'] = neutralize(industry_dummy.values.astype(float), df[factor].values).flatten()\n",
"print(df[['code', 'neutralized_factor']].head())"
]
},
{
"cell_type": "code",
"execution_count": 101,
"cell_type": "markdown",
"metadata": {},
"outputs": [],
"source": [
"df = pd.merge(df, index_components, on=['code'], how='inner').dropna()"
"## 4. 组合构建\n",
"-----------------------------\n",
"\n",
"使用alpha-mind中的线性优化器来做组合构建，在做组合构建的时候，我们已生成**主动权重**为目标（主动权重 = 组合权重 - 指数权重）。优化器，以最大化组合的预期收益为目标，同时达到以下的限制条件：\n",
"\n",
"* 最小主动权重-2%，同时保证不做空（所以单只股票最小主动权重为-2%与其指数权重负值中的较大值）；\n",
"* 最大主动权重2%；\n",
"* 行业主动暴露为0；\n",
"* 主动权重加和为0，保证组合总权重与指数总权重一致。"
]
},
{
......@@ -185,6 +281,17 @@
" np.concatenate((industry_up_bounds, [0.])),)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"输出结果中：\n",
"\n",
"* `status`：优化状态；\n",
"* `optimized_values`：组合预期收益的负值；\n",
"* `weights`：组合中股票的主动权重"
]
},
{
"cell_type": "code",
"execution_count": 112,
......@@ -208,9 +315,19 @@
" risk_target)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"计算最终持仓：\n",
"\n",
"* `portfolio_weight`：组合权重\n",
"* `active_weight`：主动权重"
]
},
{
"cell_type": "code",
"execution_count": 129,
"execution_count": 157,
"metadata": {},
"outputs": [],
"source": [
......@@ -218,18 +335,25 @@
"df['active_weight'] = weights"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"我们可以通过计算行业权重，并与指数的行业权重进行比较，验证行业暴露确实为0"
]
},
{
"cell_type": "code",
"execution_count": 137,
"execution_count": 158,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"<matplotlib.axes._subplots.AxesSubplot at 0x193c2c75f28>"
"<matplotlib.axes._subplots.AxesSubplot at 0x193c4445128>"
]
},
"execution_count": 137,
"execution_count": 158,
"metadata": {},
"output_type": "execute_result"
},
......@@ -250,6 +374,13 @@
"df.groupby('industryName1').sum().plot.bar(y=['index_weight', 'portfolio_weight'], figsize=(14, 7))"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"我们也可以通过观察中性化后因子值，观察个股权重与因子值的关系"
]
},
{
"cell_type": "code",
"execution_count": 139,
......
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