Search

Among different techniques, algorithms and applications of Data Mining, predicting the class label of unlabeled objects(undefined class label) is a crucial term in the field. The most common approaches in this area is the use of classification technique (DT, Bayes, SVM, KNN and others) that represent what is known as supervised learning. However, in many cases no target class labels and the boundaries are available to perform the prediction, so the new approach Clustering-classification technique is used. The work in this paper presents a survey of the most common researches conducted in this field and discuss their experiments, the algorithms they used, the types of data they utilized, the data sizes used, and the results they discovered. According to the results, applying the clustering techniques before classification improved classification accuracy and reduced experiment execution time. The Cluster Classifier was proven to be a suitable approach to summarize data by some of the researchers. It achieves a summarization rate of over 50%, which represents a considerable reduction in the size of the test datasets.. The findings of the researches indicated that, in addition to feature selection and feature extraction, data preprocessing (handled missing data and effective outlier detection techniques) enhanced the classifier performance and accuracy while reducing the classification error.

Clustering represents one of the most popular knowledge extraction algorithms in data mining techniques. Hierarchical and partitioning approaches are widely used in this field. Each has its own advantages, drawbacks and goals. K-means represents the most popular partitioning clusteringtechnique, however it suffers from two major drawbacks; time complexity and its sensitivity to the initial centroid values. The work in this paper presents an approach for estimating the starting initial centroids throughout three process including density based, normalization and smoothing ideas. The proposed algorithm has a strong mathematical foundation. The proposed approach was tested using a free standard data (20000 records). The results showed that the approach has better complexity and ensures the clustering convergence.

The work in this paper presents a proposed solution for preprocessing, analyzing, mining and data warehouse model for personal medical data collected from different hospitals and clinics. The proposed solution contains different phases and steps, including Extraction, Transforming and Loading (ETL) and data preprocessing focuses on converting the logged data into categories suitable for analysis and mining process, a star warehouse model was implemented that fulfills the required processing techniques, data are represented by multi-dimensional cubes for efficient and better data representation, and finally link analysis was applied on the data. The proposed framework is simple and straight forward for implementation. Personal medical data from different sources mostly in Excel files were converted into clean, complete and consistent data by different preprocessing techniques. Logged data were converted into high quality, reliable and suitable for analysis and mining process. Star warehouse schema was implemented since it is very suitable for such type of data and mining techniques. 19900 patients records were collected and used in this work. Excel and WEKA software were used for the analysis and mining processes.

Solving transportation problems where products to be supplied from one side(sources) to another (demands) with a goal to minimize the overall transportation cost represents an activity of great importance. Most of the works done in the field deals with the problem as two-sided model (Sources such as factories and Demands such as warehouses) with no connections between sources or demands. However, real world transportation problems may come in another model where sources are connected in a network like graph in which each source may supply other sources in a specific cost. The work in this paper suggests an algorithm and a graph model with mathematical solution for finding the minimum feasible solution for such widely used transportation problems. In this work, the graph representing the problem in which all sources are connected together in a network model with specific cost on each edge is converted into a new graph where additional virtual sources representing supplies between sources are added to the graph , new costs between the added sources and the demands are also calculated, and then modified Kruskal’s algorithm is applied to get the minimum feasible solution. The proposed solution is a straight forward model with strong mathematical and graph models. It can be widely used for solving real world transportation problems with feasible time and space complexity where time complexity of O(E2 + V2) is required, where E represents the number of edges and V represents the number of vertices. Different numerical examples were used to study the effectiveness and correctness of the proposed algorithm.