1. 雲端計算

2. Tensorflow
Use multiple features to further improve the effectiveness of a model
Debug issues in model input data

3. Tensorflow

4. Load the necessary libraries and import data
import math
from IPython import display
from matplotlib import cm
from matplotlib import gridspec
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
import sklearn.metrics as metrics
import tensorflow as tf
from tensorflow.python.data import Dataset
tf.logging.set_verbosity(tf.logging.ERROR)
pd.options.display.max_rows = 10
pd.options.display.float_format = '{:.1f}'.format
california_housing_dataframe = pd.read_csv(" sep=",")

5. Prepares input features from data set
def preprocess_features(california_housing_dataframe):
selected_features = california_housing_dataframe[
[“latitude”, “longitude”, “housing_median_age”, "total_rooms",
“total_bedrooms”, “population”, “households”, "median_income"]]
processed_features = selected_features.copy()
# Create a synthetic feature.
processed_features["rooms_per_person"] = (
california_housing_dataframe["total_rooms"] /
california_housing_dataframe["population"])
return processed_features
複製selected features
加入新的feature

6. Prepares target features (labels) from data set
def preprocess_targets(california_housing_dataframe):
output_targets = pd.DataFrame()
# Scale the target to be in units of thousands of dollars.
output_targets["median_house_value"] = (
california_housing_dataframe["median_house_value"] / )
return output_targets
建立空的 Daraframe
定義target feature

7. Training set
Choose the first examples, out of the total of
training_examples = preprocess_features(california_housing_dataframe.head(12000))
print(training_examples.describe())
把前12000筆資料當作training set

8. Training targets (Label)
Choose the first examples, out of the total of
training_targets = preprocess_targets(california_housing_dataframe.head(12000))
print(training_targets.describe())

9.  Validation set
Choose the last 5000 examples, out of the total of
validation_examples = preprocess_features(california_housing_dataframe.tail(5000))
print(validation_examples.describe())
把最後5000筆資料當作Validation set

10.  Validation target
Choose the last 5000 examples, out of the total of
validation_targets = preprocess_targets(california_housing_dataframe.tail(5000))
print(validation_targets.describe())

11. Plot Latitude/Longitude vs. Median House Value
plt.figure(figsize=(13, 8))
ax = plt.subplot(1, 2, 1)
ax.set_title("Validation Data")
ax.set_autoscaley_on(False)
ax.set_ylim([32, 43])
ax.set_autoscalex_on(False)
ax.set_xlim([-126, -112])
plt.scatter(validation_examples["longitude"],
validation_examples["latitude"],
cmap="coolwarm",
c=validation_targets["median_house_value"] / validation_targets["median_house_value"].max())
ax = plt.subplot(1,2,2)
ax.set_title("Training Data")
ax.set_autoscaley_on(False)
ax.set_ylim([32, 43])
ax.set_autoscalex_on(False)
ax.set_xlim([-126, -112])
plt.scatter(training_examples["longitude"],
training_examples["latitude"],
cmap="coolwarm",
c=training_targets["median_house_value"] / training_targets["median_house_value"].max())
_ = plt.plot()

12. Compare with real map

13. Find the bug
For any given feature or column, the distribution of values between the train and validation splits should be roughly equal.
The fact shows that we likely have a fault in the way that our train and validation split was created.
If we don't randomize the data properly before creating training and validation splits, then we may be in trouble if the data is given to us in some sorted order, which appears to be the case here.

14. Define input function
def my_input_fn(features, targets, batch_size=1, shuffle=True, num_epochs=None):
# Convert pandas data into a dict of np arrays.
features = {key:np.array(value) for key,value in dict(features).items()}
# Construct a dataset, and configure batching/repeating.
ds = Dataset.from_tensor_slices((features,targets)) # warning: 2GB limit
ds = ds.batch(batch_size).repeat(num_epochs)
# Shuffle the data, if specified.
if shuffle:
ds = ds.shuffle(10000)
# Return the next batch of data.
features, labels = ds.make_one_shot_iterator().get_next()
return features, labels

15. Construct the tensorFlow feature columns
def construct_feature_columns(input_features):
"""Construct the TensorFlow Feature Columns.
Args:
input_features: The names of the numerical input features to use.
Returns:
A set of feature columns
"""
return set([tf.feature_column.numeric_column(my_feature)
for my_feature in input_features])
Name of numerical input features
set():建立一個無序不重複元素的集合

16. Train model

17. 驗收 Train model

18. 驗收 Train model