Hotel-booking-demand-Case-Study: This notebook contains a series of exercises designed to explore a range of data science, python scripting, and quantitative reasoning skills. You can, in principle. solve these exercises using a number of different programming languages/environments, but it will likely be easiest for you to simply fill out this notebook with your solutions in the relevant sections following each exercise.

# Hotel-booking-demand-Case-Study
This notebook contains a series of exercises designed to explore a range of data science, python scripting, and quantitative reasoning skills. You can, in principle. solve these exercises using a number of different programming languages/environments, but it will likely be easiest for you to simply fill out this notebook with your solutions in the relevant sections following each exercise.

The data used in this case study come from the following academic work:
https://www.sciencedirect.com/science/article/pii/S2352340918315191
Feel free to peruse that paper in order to familiarize yourself with the data.

### Part 1: A First Look at the Data

#### Exercise 1

The data are in two separate csv files in this directory. Bring them into this notebook as either a
table or dataframe. Do the two datasets have the same column names (features)? If they do,
combine the data into a single table or dataframe to work with for the rest of the notebook. How
large is the dataset?

#### Exercise 2

What are the names of the features of the dataset and what are their data types (i.e. are they
numerical data? categorical? strings? If numerical, are they floats, ints, ...?)

#### Exercise 3

At this point, what other steps would you take to get a better understanding of the data? Does it
need to be cleaned or manipulated in any way? Explain and execute

### Part 2: Diving Into Some Details

Now that we have cleaned the data, put it in a friendly format, and have a basic understanding of
it, lets dive deeper into some details.

#### Exercise 1

What is the distribution of the average daily rate (ADR)? Do you notice anything peculiar about
this distribution? If so, and if you think something should be done about it, state it and do that.
Also, approximate the mean and standard deviation of this distribution by eye, and then compare
with the actual values.

#### Exercise 2

What is the distribution of values of the total length of stay? Do you notice anything noteworthy
about it?

#### Exercise 3

What is the distribution of countries which people booking stays are coming from? Report this
answer as percentages. What does this tell you about the probable location of the hotels?

#### Exercise 4

How would you investigate the relationships between the various features of the data? Explain
and demonstrate.

#### Exercise 5

Suppose we are interested in the ADR feature. Further investigate the relationship between the
ADR and some of the other features that you think might be relevant. What do you notice? Feel
free to construct new features or modify existing ones if you think it would enhance the analysis.

### Part 3: Machine Learning and Data Modeling
In this section, we will touch on some advanced data science tools to solve some realistic
problems.

#### Exercise 1

Deploy a linear machine learning model that predicts the ADR as a function of whatever single
feature you think is most strongly correlated. Are you able to get a good prediction stats on your
validation set? Why or why not?

#### Exercise 2

Deploy a machine learning model that predicts the ADR as a function of whatever set of features
you think are important. Explain your choices for the features and the type of ML model. If you
use any, how can you handle the categorical features?
How does your model perform?

#### Exercise 3

Are you happy with the performance and efficiency of your model? What could you do to improve
it? Explain and execute.

----------------------------------------------------------------------------------------

At start we have implemented linear regression and polynomial regression the results are not good. After that we tried more advance methods like Decision Tree Regressor and Random Forest Regressor. In Decision Tree Regressor the model shown quit good results then previous polynomial results but still model not perform excellent on test data set. It achieve 0.77 test score and 1.00 training data score. Because the difference quit a big we come to know that the model overfit on the data.

### Part 4: Your Turn
In this final section, you are encouraged to further demonstrate your skills and knowledge using
the provided dataset. Feel free to take things in whatever direction you find interesting.

As you can see the above parts there are number of methods but not perform state of the art. One of method of prediction is Random forest regression. It has some advantage of dealing with over fit. In it we choose number of trees that is directly made effect of regularization.  If the less you chose tress the more model will go towards linearization.  You can change tress number n_estimators in the code to check the different result of accuracy. It perform good then other methods.

All these methods perform different in different problems here a some over view of these methods

![Part 4](part_4.jpg)

yanhouzhen/Hotel-booking-demand-Case-Study

简介

发行版

贡献者

近期动态

yanhouzhen/Hotel-booking-demand-Case-Study .gitee-modal { width: 500px !important; }

简介

发行版

贡献者

近期动态

搜索帮助

yanhouzhen/Hotel-booking-demand-Case-Study