利用TensorFlow进行高效预测分析

需积分: 10 128 浏览量更新于2024-07-19 收藏 2.94MB PDF 举报

"TensorFlow_Powerful Predictive Analytics with TensorFlow-Packt Publishing2018" 在当前大数据时代，预测分析已经成为商业智能中的关键工具，它通过挖掘结构化和非结构化数据中的隐藏模式来实现自动化决策。这本书"TensorFlow: Powerful Predictive Analytics with TensorFlow"由Md. Rezaul Karim撰写，旨在引导读者掌握使用TensorFlow进行监督学习、无监督学习和强化学习的方法。 TensorFlow是一个强大的开源库，由Google Brain团队开发，主要用于机器学习和深度学习。它支持构建和部署复杂的计算模型，尤其适合处理大规模数据流和实施预测分析。书中详细介绍了如何利用TensorFlow的特性来处理各种类型的预测任务。监督学习是机器学习的一个分支，它依赖于带有标签的数据来训练模型，从而对未知数据进行预测。例如，通过监督学习，我们可以构建一个模型，该模型可以根据历史销售数据预测未来的销售额。TensorFlow提供了丰富的API，如神经网络层和优化器，使得构建这样的模型变得简单易行。无监督学习则不依赖于标记数据，而是试图发现数据中的内在结构和模式。例如，聚类算法可以用来识别用户群体，这些群体可能具有相似的购买行为或偏好。TensorFlow可以与Keras等高级API结合使用，实现这些无监督学习任务。强化学习则侧重于通过与环境的交互来学习最优策略。比如，AI游戏代理通过不断尝试和学习，以最大化奖励。在TensorFlow中，可以使用Q-learning或策略梯度等算法来实现强化学习模型。此外，书中还可能涵盖了数据预处理、特征工程、模型评估和调优等核心概念。数据预处理是机器学习流程的关键步骤，包括数据清洗、标准化和编码。特征工程涉及选择最有影响力的特征以及转换数据以提高模型性能。模型评估通常涉及交叉验证和各种性能指标（如准确率、精确率、召回率和F1分数），而调优则是通过调整超参数来优化模型的性能。通过深入学习TensorFlow，读者将能够利用机器学习的力量来解决实际的预测问题，无论是在金融、医疗、零售还是其他行业。书中提供的实践案例和代码示例将帮助读者将理论知识转化为实际应用，从而在海量信息中挖掘出有价值的洞见并作出明智的决策。这本书为那些希望掌握预测分析和TensorFlow的读者提供了一个全面且实践性强的学习资源。

Lesson 1

[ 5 ]

The ultimate target of using this dataset is to predict what kind of people survived

the Titanic disaster. However, a bit of exploratory analysis of the dataset is a

mandate.Atrst,weneedtoimportnecessarypackagesandlibraries:

import pandas as pd

import matplotlib.pyplot as plt

import numpy as np

Now read the dataset and create a panda's DataFrame:

df = pd.read_csv('/home/asif/titanic_data.csv')

Before drawing the distribution of the dataset, let's specify the parameters for

the graph:

fig = plt.figure(figsize=(18,6), dpi=1600)

alpha=alpha_scatterplot = 0.2

alpha_bar_chart = 0.55

fig = plt.figure()

ax = fig.add_subplot(111)

Draw a bar diagram for showing who survived versus who did not:

ax1 = plt.subplot2grid((2,3),(0,0))

ax1.set_xlim(-1, 2)

df.Survived.value_counts().plot(kind='bar', alpha=alpha_bar_chart)

plt.title("Survival distribution: 1 = survived")

Plot a graph showing survival by Age:

plt.subplot2grid((2,3),(0,1))

plt.scatter(df.Survived, df.Age, alpha=alpha_scatterplot)

plt.ylabel("Age")

plt.grid(b=True, which='major', axis='y')

plt.title("Survival by Age: 1 = survived")

Plot a graph showing distribution of the passengers classes:

ax3 = plt.subplot2grid((2,3),(0,2))

df.Pclass.value_counts().plot(kind="barh", alpha=alpha_bar_chart)

ax3.set_ylim(-1, len(df.Pclass.value_counts()))

plt.title("Class dist. of the passengers")

From Data to Decisions – Getting Started with TensorFlow

[ 8 ]

Well, we have seen the data. Now it's your turn to do some analytics on top of the

data. Say predicting what kinds of people survived from that disaster. Don't you

agree that we have enough information about the passengers, but how could we

do the predictive modeling so that we can draw some fairly straightforward

conclusions from this data?

For example, say being a woman, being in 1st class, and being a child were all

factors that could boost passenger chances of survival during this disaster.

In a brute-force approach–for example, using if/else statements with some sort

of weighted scoring system, you could write a program to predict whether a given

passenger would survive the disaster. However, does writing such a program in

Pythonmakemuchsense?Naturally,itwouldbeverytedioustowrite,difcultto

generalize,andwouldrequireextensivenetuningforeachvariableandsamples

(thatis,passenger).

This is where predictive analytics with machine learning algorithms and emerging

tools comes in so that you could build a program that learns from the sample

data to predict whether a given passenger would survive. In such cases, we will

see throughout this book that TensorFlow could be a perfect solution to achieve

outstanding accuracies across your predictive models. We will start describing the

general overview of the TensorFlow framework. Then we will show how to install

andcongureTensorFlowonLinux,MacOSandWindows.

General Overview of TensorFlow

TensorFlowisanopensourceframeworkfromGoogleforscienticandnumerical

computationbasedondataowgraphsthatstandfortheTensorFlow'sexecution

model.ThedataowgraphsusedinTensorFlowhelpthemachinelearningexperts

to perform more advanced and intensive training on the data for developing

deep learning and predictive analytics models. In 2015, Google open sourced the

TensorFlow and all of its reference implementation and made all the source code

available on GitHub under the Apache 2.0 license. Since then, TensorFlow has

achieved wide adoption from academia and research to the industry, and following

thatrecentlythemoststableversion1.xhasbeenreleasedwithauniedAPI.

As the name TensorFlow implies, operations are performed by neural networks

onmultidimensionaldataarrays(akaowoftensors).Thisway,TensorFlow

provides some widely used and robust implementation linear models and deep

learning algorithms.

Lesson 1

[ 9 ]

Deploying a predictive or general purpose model using TensorFlow is pretty

straightforward. The thing is that once you have constructed your neural networks

model after necessary feature engineering, you can simply perform the training

interactivelyusingplottingorTensorBoard(wewillseemoreonitinupcoming

sections). Finally, you deploy it eventually after evaluating it by feeding it some

test data.

Sincewearetalkingaboutthedataowgraphs,nodesinaowgraphcorrespond

to the mathematical operations, such as addition, multiplication, matrix factorization,

and so on, whereas, edges correspond to tensors that ensure communication between

edgesandnodes,thatisdataowandcontrolow.

You can perform the numerical computation on a CPU. Nevertheless, using

TensorFlow, it is also possible to distribute the training across multiple devices

on the same system and train on them, especially if you have more than one GPU

on your system so that these can share the computational load. But the precondition

is if TensorFlow can access these devices, it will automatically distribute the

computations to the multiple devices via a greedy process. But TensorFlow also

allows the program, to specify which operations will be on which devices via name

scope placement.

The APIs in TensorFlow 1.x have changed in ways that are not all backward

compatible. That is, TensorFlow programs that worked on TensorFlow 0.x won't

necessarily work on TensorFlow 1.x.

The main features offered by the latest release of TensorFlow are:

• Faster computing: The latest release of TensorFlow is incredibly faster.

For example, it is 7.3 times faster on 8 GPUs for Inception v3 and 58 times

speedupfordistributedinception(v3trainingon64GPUs).

• Flexibility: TensorFlow is not just a deep learning library, but it comes

with almost everything you need for powerful mathematical operations

through functions for solving the most difficult problems. TensorFlow 1.x

introduces some high-level APIs for high-dimensional arrays or tensors, with

tf.layers, tf.metrics, tf.losses, and tf.keras modules. These have

made TensorFlow very suitable for high-level neural networks computing.

• Portability: TensorFlow runs on Windows, Linux, and Mac machines and on

mobilecomputingplatforms(thatis,Android).

• Easy debugging: TensorFlow provides the TensorBoard tool for the analysis

of the developed models.

• Unified API: TensorFlow offers you a very flexible architecture that enables

you to deploy computation to one or more CPUs or GPUs in a desktop,

server, or mobile device with a single API.

剩余164页未读，继续阅读

WindStand

粉丝: 35

利用TensorFlow进行高效预测分析

最新资源