Annotated Bibliography on Data Mining

Annotated Bibliography on Data Mining

Thesis: Data handlers have benefitted immensely from data mining and should be allowed to continue with their activities as long as they are legal.

Rao, Rajinder Singh, and Jyoti Arora. “A survey on methods used in web usage mining.” Int.        Res. J. Eng. Technol 4.5 (2017): 2627-2631.

The information found on the World Wide Web is rich with the knowledge required for most of the challenges facing the world. However, obtaining information from the internet is not an easy task, and data handlers are needed to help in the required access for the information. Rao and Arora surveyed the methods used in web usage mining where data handlers employ data mining techniques to avail data for potential users from the internet. The article combines the implementation of traditional data mining methods to embrace technology to obtain information from the internet. The researchers identify three classes of knowledge developed through data mining: Web activity obtained from server logs and Web browser activity tracking, Web graph that is derived from the links between the pages and Web content that includes the data found on Web pages and inside of documents. The primary theme in the article is web usage that relies on data mining to provide the information required on the internet.

The demand for information came along with the need for more websites that are aimed at providing the required information. However, the information on the websites is not sorted to fit into the requirements of the information users. Web usage comes in to aid in the filtering of information to fit into the requirements of the users. The move to sort the information on the websites to meet the needs of the users creates different phases that are crucial in the whole process. The authors conclude that data mining is essential in the era of large data, and data handlers should be allowed to carry out their activities. Besides, data would of little or no use if the target users are not allowed to access the information in the desired form.

Sengupta, Shantashree. “Applications of Data Mining in Library & Information Centres: An             Overview.” International Journal of Current Research 9.01 (2017): 45246-45249.

The increase in the amount of data to process has posed a significant challenge to libraries and information centers. In 2017, Sengupta researched to assess the application of data mining in the library and information centers. The problems posed b dealing with large data sets have led to the need to develop mechanics of identifying patterns in the samples, and data mining is one of the mechanics. The authors assessed the application of data mining to statistics, machine learning, and artificial intelligence and database systems. Furthermore, the research addresses data mining techniques that are essential in creating meaningful information from the information found on websites. According to the article, data mining involves looking at patterns in more massive data sets that have not been previously identified to add meaning to the conclusions developed from the datasets. Generally, the article presents various techniques used in data mining and gives the advantages and disadvantages of data mining.

Data mining is presented as having several techniques that aid in the identification of patterns and hidden information on datasets. The association technique is, for example, defined as being applicable in the marketing field to assess the trends existing between items within similar transactions. Another method that is identified in the article is the classification technique that is said to be relevant for machine learning. The clustering technique adds sense to the classification technique by grouping items according to the clusters of objects that are identified in classification. Generally, the article classifies the techniques according to their fields of application and shows that some techniques can be used collaboratively. The researchers conclude that the application of data mining is broad and cannot be overlooked in the field of library and information centers.

Dabhi, Dipak P., and Mihir R. Patel. “Extensive survey on hierarchical clustering methods in         data mining.” International Research Journal of Engineering and Technology (IRJET) 3            (2016): 659-665.

Data mining has appeared to be a combination of several techniques that are all aimed at bringing out hidden patterns on the information obtained from websites. Dabhi and Patel researched on the clustering method used in data mining and came up with several conclusions. The clustering approach employs several techniques such as Partitioning, Hierarchical, density-based, model-based, and grid-based methods with each method offering a set of advantages and disadvantages. However, no clustering method has proved to provide solutions to all the challenges posed by bid data analysis. The researchers base their argument on the hierarchical clustering method, which considering grouping items in datasets from n number of items to fewer groups of items with similar characteristics. The authors study the various methods of hierarchical clustering and summarize the details of each technique. Also, the authors provide details concerning the various clustering methods and explain the advantages and disadvantages associated with every clustering method. Generally, the article provides a quick review of the different methods of hierarchical clustering as used in data mining. Researchers willing to study data mining as a way of economics will find information from the article essential.

Dutt, Ashish, Maizatul Akmar Ismail, and Tutut Herawan. “A systematic review on educational   data mining.” IEEE Access 5 (2017): 15991-16005.

Another application of data mining has been in the educational sector. Data handlers have been in demand to help in the challenges that have been facing educational institutions while addressing the issue of student data. Generally, educational institutions are forced to store large data sets that carry information about the students, and analyzing such data becomes a challenge. Besides, institutions end up losing significant patterns that can only b found through data mining. Dutt, Ismail, and Herawan did a study in 2017 to assess the implementation of data mining in educational sectors. The researchers present a review of a systematic clustering algorithm that has been in use for over three decades. According to the literature reviewed by the researchers, several recommendations are made concerning possible improvements in the future of data mining in the field of education. Furthermore, the researchers outline the advantage that clustering brings out the students’ study patterns by considering various variables that would be overlooked if data mining was not employed. Also, clustering ensures that only the relevant variables are used in the data mining process to ensure that the patterns are adequately addressed. Generally, the researchers argue that educational data is multi-level, and thus researchers should always make informed decisions concerning the right clustering technique to use in their studies.

Yee, Ong Shu, Saravanan Sagadevan, and Nurul Hashimah Ahamed Hassain Malim. “Credit        card fraud detection using machine learning as data mining technique.” Journal of       Telecommunication, Electronic and Computer Engineering (JTEC) 10.1-4 (2018): 23-27.

Cases of credit card fraud have been on the rise owing to the technological developments in the financial sectors. Criminals have been using information from websites to carry out their fraudulent activities with or without the knowledge of their victims. The fraudulent activities have forced several authorities to come up with strategies to detect the activities and take the necessary actions before the consequences fall out of hand. Yee, Sagadevan, and Malim did a study in 2018 to assess the efficiency of machine learning in the creation of detection methods for credit card fraud. According to the researchers, the cases of credit card fraud could be associated with the rising number of online card transactions that expose customers’ information to fraudsters; hence, the increasing rate of crime. The stud employed the use of Bayesian metrics to assess the trend of fraudulent cases as well as the efforts made by various authorities to address the growing concern. Furthermore, two datasets were analyzed in the study to come up with the conclusions and recommendations made. The datasets were customized to have primary data and secondary data that were filtered to eliminate the loopholes found in the first dataset. Generally, the researchers concluded that filtered data was safer compared to the unfiltered data that is held by credit card companies.