Regression Algorithm for Business Analysis in Boston Housing Authority

This report discusses the implementation process of regression in business management, focusing on the Boston Housing Authority. It explores the business analysis problems faced by the authority and the benefits of using linear regression algorithm. The report provides insights from the results and suggestions for business management.

Running head: REGRESSION ALGORITHM
REGRESSION ALGORITHM
Name of the Student:
Name of the University:
Author Note:

1REGRESSION ALGORITHM
Table of Contents
Introduction......................................................................................................................................2
Business analysis problem...............................................................................................................3
Methodology....................................................................................................................................3
Result and Discussion......................................................................................................................6
Conclusion.......................................................................................................................................8
References......................................................................................................................................10

2REGRESSION ALGORITHM
Introduction
Background
Regression analysis have been providing high benefits to the business organization due to
its ability to perform predictive analysis. Hence with the help of predictive analysis the main
advantages for the Boston Housing Authority. BHA is considered as a public agency as they
intend on providing public housing at a lower cost. The cost incurred is suitable for the low and
mediocre income drawing families. A total of 26000 people globally has been served by the
equipped functionality of Boston housing authority. This business hub was established in the
year 1935. It is found that linear regression analysis has a huge effect on the functioning of the
BHA.
Project objective
The objective of the project is to get well accustomed with the implementation process of
the regression in the field of business management. Another major objective of the project is to
understand the vivid sets of algorithms and methods that can be implemented in business
management sectors. Getting introduced to the algorithms of regressions that are used in the data
mining processes are also the major objective of the project
Overview of the report
This report will be discussing about the business analysis problems that are faced by
BHA. This report also discusses about the business analysis tasks and their aims. Benefits that
will be received after implementation of the task is also stated. The mining methods and
algorithms that are used for resolving the issues are also stated in the process. A brief discussion

3REGRESSION ALGORITHM
regarding the algorithm that is used is also stated in the report. A brief introduction to the data set
is also made in this report. Preprocessing of the data set is also performed. The results that are
gained are also well discussed in the report. Recommendations are also provided in the report.
Business analysis problem
Business analysis problem
The main issues that are present in the business analysis of the BHA includes improper
processing of the portfolio management of the buildings. The rent that is paid by the residents of
the BHA are huge and hence they face difficulties in paying the rent.
Aim of Business analysis tasks
The aim of the business analysis tasks will be ton provide a better understanding of the
issues that are present in the project management system. This incurs the fact that better
management of the projects that are undertaken by BHA will get better and hence efficiency in
the business will increase.
Benefits
The benefits received will be for the clients and hence they will stay content and hence
retention of the clients will be a main benefit. This will also ensure that the revenue that will be
earned will be having a benefitting BHA itself (Isikhan et al, 2019).
Methodology
Data mining algorithm used
Usage of Linear regression algorithm is used in this case. Linear regression algorithm
integrated with data mining and machine learning has been the major aspect that facilitates the

4REGRESSION ALGORITHM
functional process. This inclusion of the linear regression algorithm the main concern that is
present in the discussion regarding gradient and descent (Ueno and Nagata 2017). Linear
regression is a statistical analysis of the attempts that shows the relationship in between the
variables that are present. Linear regression promotes the processing of various data and hence
serving the trend lines. Linear regression is performed as per the plotting of the trends and hence
this process helps in proper prediction of the output. This process is performed based on the
apparent random data and hence properly diagnosing the stock price.
Reason of selection
Linear regression algorithm is used for establishing a relation in between 2 continuous
variables. It is a machine learning algorithm that enhances proper management of the supervised
regression task. Differentiating in between the target prediction that is value based and the
independent variables are the major utility of the algorithm (Muralidharan et al, 2018). Linear
regression is performed in order to predict the value of the dependent variable on the
independent variable. The functions that are important includes cost function that is described by
J. Implementation of the best fit regression line, prediction of the independent value is
performed. This has been a major issue due to which the importance of the functioning process is
performed. The error in between the predicted value and the true value is calculated (Zhang, Lian
and Yu 2017). It is desired that the error in between the two values are minimal. Hence updating
of the intercept and the coefficient values are important. In this case the main advantage that will
be received includes proper management of the root mean square of the values in between the
predicted value and the true value (Abid and Zou 2018). With proper updating of the value the
main advantage that is gained includes error in between the predicted value and the real value

5REGRESSION ALGORITHM
decreases. Gradient descent is concerned with the updating of the values of the intercept and the
gradient in order to reduce the cost function. This also ensures that RMSE value gets decreased.
The benefits of using this model are as follows: -
The main benefit that is received includes predicted outcome of the project. With
predicted outcome of the project the risks that might be present will be well analyzed and hence
proper management of the data sets will be performed. Understanding of the inventory levels
will also be beneficial in the process (Kondo, Hayashi and Maeda 2017). Understanding of the
supply and demand of the project will also be beneficial. Data analysis regarding the reasons of
output generated in a gradual interval of time helps in better management of the system. Proper
prediction of sales in near future will also be beneficial for the business organization that will be
implementing the same (Her et al, 2018).
What data set is to be used?
Boston Housing data set is used in the project.
Data cleaning
The steps that are followed for implementing the data cleaning process includes removing
of the unwanted observation. This section has been performing the management of the removal
of unwanted observations that includes observations that have been duplicated and irrelevant
observations (Zambom and Ravikumarr 2017). Duplicate observation includes combining of
datasets from different places and hence excess data storage will be present. Scrape data set is
also present. This includes the fact that better management of the data duplication issue can be
performed. Cleaning the data also includes getting free from the data that are out of specification
of the considered data process (Ben-Porat and Tennenholtz 2017).

6REGRESSION ALGORITHM
Result and Discussion
Research Question
Research Question 1: find out the percentage of the owner occupied homes in comparison to the
non-retail business acres and residential lands?
Research Question 2: What are the factors that affect the percentage of the owners occupying
homes in Boston?
Present your results and findings
SUMMA
RY
OUTPUT
Regression Statistics
Multiple R 0.73766
2726
R Square 0.54414
6298
Adjusted
R Square
0.54324
1826
Standard
Error
6.21576
0405
Observatio
ns
506

7REGRESSION ALGORITHM
ANOVA
df SS MS F Signific
ance F
Regressio
n
1 23243.914 2324
3.91
601.6
179
0.0000
Residual 504 19472.381
42
38.63
568
Total 505 42716.295
42
Coeffici
ents
Standard
Error
t Stat P-
value
Lower
95%
Upper
95%
Lower
95.0%
Upper
95.0%
Intercept 34.5538
4088
0.5626273
55
61.41
515
3.7E-
236
33.4484
5704
35.6592
2472
33.4484
5704
35.65922
472
lstat -
0.95004
9354
0.0387334
16
-
24.52
79
5.08
E-88
-
1.02614
82
-
0.87395
051
-
1.02614
82
-
0.873950
508
Discuss your results
The variables that are used in the regression statistics includes Multiple R, R Square,
Adjusted R Square, Standard Error and Observations. The percentage that is observed in the R
Square is 54.41%. After completion of the Regression calculation, it is seen that the entire
prediction is highly significant. As per the regression statistics, the value of Multiple R is

8REGRESSION ALGORITHM
0.737663. The value of Adjusted R Square that is considered includes 0.543242 and the value of
the Standard error is 6.2157 and the observation is found to be 506. Usage of linear regression is
made, hence it can be stayed that the processing of the business management will be performed
and hence commencement of the Intercept and lstat has coefficients of 34.5568 and -0.95004
respectively. The standard value set present includes Intercept and lstat is 0.562627 and
0.038733 respectively. The t stat value generated is 61.4151455 and -24.5278999 respectively.
Intercept and lstat value of P value is 3.7431E-236 and 5.08E-88 respectively.
Provide any suggestion to business management
 The suggestions that can be provided includes proper management of the tax rates that
are included in the system.
 Decrease in the cost of the projects are also expected
 Proper understanding of the requirement of the people residing can be provided.
Conclusion
From the above discussion it can be stated that Linear regression have been implemented
for the processing of the Boston housing dataset. This report contains different sections of houses
that are present in Boston. This is the major aspect that proper management of the business data
set will be provided with data regarding the UCI machine learning and data mining processes. A
total of 506 samples and 13 features are provided and hence predictive analysis is performed.
With the help of the predictive data analysis the main advantage received includes better
processing of the business management. The significance of the project data collected is
considered to be a major benefit as per the consideration of the BHA system.

9REGRESSION ALGORITHM
References
Abid, A. and Zou, J., 2018. Stochastic em for shuffled linear regression. arXiv preprint
arXiv:1804.00681.
Ben-Porat, O. and Tennenholtz, M., 2017. Best response regression. In Advances in Neural
Information Processing Systems (pp. 1499-1508).
Her, Q.L., Vilk, Y., Young, J., Zhang, Z., Malenfant, J.M., Malek, S. and Toh, S., 2018. A
distributed regression analysis application based on SAS software. Part I: Linear and logistic
regression. arXiv preprint arXiv:1808.02387.
Isikhan, S.Y., Karabulut, E., Samadi, A. and Kılıçkap, S., 2019. Adaptation of Error Adjusted
Bagging Method for Prediction. International Journal of Data Warehousing and Mining
(IJDWM), 15(3), pp.28-45.
Kondo, Y., Hayashi, K. and Maeda, S.I., 2017. Sparse Bayesian linear regression with latent
masking variables. Neurocomputing, 258, pp.3-12.
Muralidharan, S., Phiri, K., Sinha, S.K. and Kim, B., 2018. ANALYSIS AND PREDICTION OF
REAL ESTATE PRICES: A CASE OF THE BOSTON HOUSING MARKET. Issues in
Information Systems, 19(2), pp.109-118.
Ueno, Y. and Nagata, Y., 2017. A Proposal of Regression Hybrid Modeling for Combining
Random Forest and X-Means Methods. Total Quality Science, 3(1), pp.1-10.
Zambom, A.Z. and Ravikumar, P., 2017. A Comparative Study of Additive and Nonparametric
Regression Estimators and Variable Selection Procedures. World Academy of Science,

10REGRESSION ALGORITHM
Engineering and Technology, International Journal of Mathematical, Computational, Physical,
Electrical and Computer Engineering, 11(3), pp.109-115.
Zhang, Y., Lian, H. and Yu, Y., 2017. Estimation and variable selection for quantile partially
linear single-index models. Journal of Multivariate Analysis, 162, pp.215-234.