STUDENT NAME:
REGISTRATION NUMBER:
COURSE CODE: CIS8008
COURSE NAME: BUSINESS INTELLIGENCE
PART B
Question One: Decision support systems have evolved into data analytics and now play a key role in supporting data-driven decision making in the big data era and the Fourth Industrial Revolution. Identify and discuss three key drivers of the Fourth Industrial Revolution underpinned by data analytics and machine learning.
The three key factors that have contributed to the fourth and most recent industrial revolution include;
- Reduced Cost of Technical Devices and Power Used in Computing
This factor has made it possible for most organizations to integrate their business systems with recent technological advancements for enhanced decision-making. Data analytics has played a significant role in the reduced cost of technical devices and computing power. This is because techniques in data analysis, such as qualitative research, form the foundation of recent technological innovations that have led to the birth of more cost-efficient devices used in computing.
- Ease of Utilizing Techniques of Artificial Intelligence and Machine Learning Algorithms
Since the introduction of Artificial Intelligence, there have been significant improvements in machine-learning algorithms, such as the simplification of syntax, which has led to more natural understanding and use of these algorithms by programmers. These improvements have made it possible for the implementation of machine learning in most business systems, making them autonomous and improving decision-making practices within an organization.
- Enhanced Availability of Cloud Computing Solutions
Data analytics has led to the realization of an explosion in data in most organizations as a result of incorporating technology in business systems. This data explosion increased the need for organizations to store their data in the cloud, which is more scalable and secure. The increase in cloud computing solutions has led to enhanced availability and reduced price, making it easier for more organizations to utilize cloud computing services.
Question Two: Identify and describe the four main steps in data pre-processing. For each of these four main steps of data pre-processing, identify and discuss one key method used.
The four main steps involved in pre-processing of data include;
- Data Consolidation
Relevant data is gathered from different sources and stored in a centralized location using data warehouses. The collected data is then evaluated to gain a better understanding that will aid in choosing the necessary information. Multiple sources of data could provide similar information leading to an unreliable data model. The data virtualization technique is used in this step to give a real-time observation of data coming in from different sources to integrate similar data.
- Data Scrubbing
Missing data values and noisy data account for some of the data anomalies that a data analyst could encounter. Missing data values can be fixed using regression or decision trees to predict the most probable values that can be filled in to replace the missing value. Regression techniques such as linear regression can also be used to smooth noisy data by assigning values to functions.
- Data Transformation
This step involves tailoring the data for better processing. One of the essential methods in this step is data normalization whereby the range in data is between the maximum value, and the minimum value is scaled down to a smaller range, for instance, 0-1. This is to prevent prominence of some data sets with large values such as monthly salary over other data sets with lower values such as years of experience.
- Reduction of Data
This step entails reducing large data values to a more manageable size. The primary purpose of this step is to enhance efficiency while dealing with the data. One method that can be used to reduce the size of large data values is doing away with attributes of the data that have more missing values than is allowed for the data model.
Question Three: The following confusion matrix was generated to evaluate the performance of a classification model using a Decision Tree for predicting whether a patient has Alzheimer’s disease or not. YES being True, a patient has Alzheimer’s disease; and NO being False, a patient does not have Alzheimer’s disease. Using this Confusion Matrix:
| Predicted | Predicted | ||
| No | Yes | ||
| 37 | 170 | Yes | Actual |
| 105 | 36 | No | Actual |
- a) State the formula used, calculate and discuss what the Accuracy Rate tell us about the
above model for predicting if a patient has Alzheimer’s disease.
The formula for the accuracy metric is as follows;
According to this model;
TP = 170
TN = 105
FP = 36
FN = 37
Therefore, the accuracy rate of this model is as follows;
The accuracy rate of this model (0.79) indicates that the accuracy of this model in predicting whether a patient has Alzheimer’s disease or not is reliable. The accuracy rate is closer to 1.0, which represents complete accuracy.
b) State the formula used, calculate and discuss what the Sensitivity Rate tells us about the
above model for predicting if a patient has Alzheimer’s disease.
Sensitivity rate is determined using the following formula;
Therefore, the sensitivity rate of this formula is as follows;
The calculated sensitivity rate of this model is high, indicating that this model is effective in predicting whether a patient has Alzheimer’s or not. This is because the high sensitivity rate of a test model shows that it often gives correct predictions.
c) State the formula used, calculate and discuss what the Specificity Rate tells us about the
above model for predicting if a patient has Alzheimer’s disease.
Specificity rate is determined using the following formula;
Therefore, the specificity rate of this formula is as follows;
The specificity rate of this model is high since it is close to 1.0, which indicates that this model effectively predicts healthy individuals because it often correctly predicts that healthy patients do not have Alzheimer’s disease.
d) State the formula used, calculate and discuss what the F1 Score tells us about above
model for predicting if a patient has Alzheimer’s disease.
The formula for calculating F1 Score of this model is as follows;
The formula for calculating precision is;
Therefore, the precision of this model is;
The formula for calculating recall is;
Therefore, the recall of this model is;
Therefore, the F1 Score of this model is;
The F1 Score for this model is high and close to 1, representing both perfect precision and recall of the model. This high F1 Score means that the model gives the most relevant prediction results while testing patients for Alzheimer’s disease.
Question Four: Describe the term Social Media Analytics and discuss the impact social media sites like Facebook and Twitter can have on customer sentiment and how this can be measured using social media analytics.
The term Social Media Analytics refers to the ability of an organization to enhance its competitiveness by utilizing resources such as tools and techniques available on social media sources based on the web through scientific and systematic means. Customer sentiment refers to a customer’s emotions or feelings regarding an organization’s product or service. Social media has become an important channel through which organizations communicate with their customers or market their products. This has a significant impact on the customer’s perceived emotions or feelings towards the product or service being sold. Social media could elicit either positive or negative customer sentiments based on numerous factors such as feedback or comments from other users of the product or service registered on the same social media platform.
The following techniques can be used to measure the impact social media has on customer sentiment;
- Descriptive analytics – This technique utilizes simple methods to identify the various activities on a social media platform, such as the number of followers gained or the amount of feedback generated. This tool can also be used to identify the social media platform used by most of the organization’s customers.
- Social network analysis – This technique analyses the online friends and followers of an organization’s customers to understand the existing links and relationships. This understanding helps identify the levels of influence between friends and followers on a social media platform.
- Advanced analytics – This technique is used to conduct a more in-depth investigation by analyzing conversations on social media platforms to identify themes and topics of discussion that would help predict customer preferences and sentiments on a future product or service.
Question Five: Describe what is meant by the term Stream Analytics and provide an example of how to stream analytics is being used in a real-world situation. Discuss two key benefits that are being realized in your example.
Stream analytics can also be referred to as data-in-motion analytics or real-time data analytics. It is used to describe action-oriented information gathered from streams of data. Stream Analytics has a variety of areas in which it is applied in the real world, and immense benefits are obtained from it. For instance, one of the sectors in which stream analytics has been implemented is in the energy industry. There has been a recent development in the energy industry whereby smart power grid systems have been formulated to be used in the supply of electric power. Since power distribution is a complex process, vast amounts of data are involved, and this brings about the need for an effective system that will best handle this process to meet customer satisfaction.
The implementation of stream analytics in the energy industry, particularly in the maintenance of smart power grid systems, has led to their enhancement in that, they are now able to utilize real-time methodologies in processing the enormous amounts of information involved in power distribution. This processing of data in real-time makes it possible to determine the most effective and optimal methods that can be used to distribute power, which will eventually lead to meeting the needs of customers and, ultimately, customer satisfaction. Power distribution also grapples with the issue of customer needs and demands arising unexpectedly. The implementation of stream analytics in the new smart power grid systems makes it possible to come up with relevant short term predictions that shed light on some of these unexpected demands as well as peaks in the generation of renewable energy.
Question Six: Describe the four major aspects of Internet of Things (IoT) Technology Infrastructure and explain how IoT technology could be used in a hospital ward to improve and complement patient care with two specific examples
The four significant aspects of the Internet of Things (IoT) technology infrastructure include;
- Hardware
This feature entails the various physical devices used to facilitate the Internet of Things, such as monitors and sensors used in IoT devices such as temperature sensors and actuators used to generate and store data. Controlling, monitoring or tracking is only necessary for the physical devices, whereas sensor devices could utilize a computing device or any other type of processor to process input data.
- Connectivity
A hub or base station should be established to gather the information from obtained from the sensor-equipped devices and transmit that information to the cloud for efficient and safe storing. Since devices within an organization need to communicate with each other or with the applications running on them, they need a reliable network connection. This means that the devices might be connected to the internet, either directly or indirectly. Indirectly connected devices require a gateway to access the data stored in the cloud.
- Software backend
Once the collected data is successfully stored, it needs to be managed. The management of gathered and stored data takes place in this aspect of IoT infrastructure. This aspect is responsible for maintaining the connected devices and the network being used to keep these devices connected. Furthermore, this aspect of IoT infrastructure facilitates the integration of collected information even while stored in the cloud.
- Applications
In this phase of IoT infrastructure, the collected and stored data is transformed into meaningful information and can be used to produce deliverables. A vast number of applications run on computing devices such as smartphones and personal computers. In contrast, others run and rely on a server to generate deliverables to stakeholders via message alerts.
IoT can be beneficial in the healthcare sector, whereby some of the IoT devices used in the hospital ward can be fitted with sensors to keep track of various hospital equipment such as wheelchairs. Furthermore, IoT devices can be connected to a network within the hospital to facilitate remote monitoring of patients by doctors.
Question Seven: Analytics and Artificial Intelligence (AI) rely on identifying patterns in large amounts of data to accurately predict outcomes, which are often used in automation of operational decision making. However exceptional events like the recent bush fires in Australia and globally the COVID-19 Pandemic can play havoc with the accuracy and reliability of algorithms underpinning analytics and data-driven decision-making. Discuss how these analytics algorithms should be evaluated and made more reliable and robust in dealing with exceptional events, changing contexts, and norms.
The following means can be used to evaluate algorithms and enhance them to deal with unexpected events, changing contexts and norms;
- Evaluating the data used in analytics algorithms to identify existing patterns
Most analytics algorithms repeatedly use the same data to produce deliverables. The best way to analyze data used in algorithms is through various machines, computing devices, and business systems. This is because these machines and systems often execute similar processes repeatedly. For instance, if a thermal sensor reports a bush fire every month due to climate change and suddenly begins to report these possible occurrences every month, the analytics algorithm should raise a flag on unusual circumstances that require investigation.
- By using accurate samples while modeling human behavior to evaluate an algorithm
When analyzing human beings, it is recommended to avoid being too hasty and not make unnecessary assumptions while gathering data and selecting the sources of your data. To come up with a useful algorithm data sources that are more inclusive should be used to reduce the chances of missing an aspect that is crucial in the event, some different type of data was used.
- Iterating and frequently improving analytics algorithms
While developing an algorithm, the first algorithm developed is a result of various assumptions and hypotheses made. It is not definite whether the algorithm will produce the expected results or not. To meet all the stakeholders’ expectations, the algorithm should be continuously run, and improvements are implemented in the areas where the algorithm does not produce the expected deliverables. This process of continuously running an algorithm is referred to as iteration.