Sequential Clustering for Event Sequences and Its Impact on Next Process Step Prediction

Abstract

Abstract. Next step prediction is an important problem in process analytics and it can be used in process monitoring to preempt failure in business processes. We are using logfiles from a workflow system that record the sequential execution of business processes. Each process execution results in a timestamped event. The main issue of analysing such event sequences is that they can be very diverse. Models that can effectively handle diverse sequences without losing the sequential nature of the data are desired. We propose an approach which clusters event sequences. Each cluster consists of similar sequences and the challenge is to identify a similarity measure that can cope with the sequential nature of the data. After clustering we build individual predictive models for each group. This strategy addresses both the sequential and diverse characteristics of our data. We first employ K-means and extent it into a categorical-sequential clustering algorithm by combining it with sequential alignment. Finally, we treat each resulting cluster by building individual Markov models of different orders, expecting that the representative characteristics of each cluster are captured.

Data mining for statistical analysis of money laundering transactions

“…There are two types of partitional clusters: (1) sequential, and (2) simultaneous. Sequential clustering uses algorithm to separate the data points in particular order (Le et al , 2014). The order of the objects then influenced the cluster that they will be partitioned into.…”

Explainable Artificial Intelligence for Process Mining: A General Overview and Application of a Novel Local Explanation Approach for Predictive Process Monitoring

“…─ next event prediction [6,[11][12][13][14][15][16][17][18] ─ business process outcome prediction [19][20][21][22] ─ prediction of service level agreement violations [15,23] There are also various studies that handle various regression problems in the business process prediction domain:…”

“…─ remaining time prediction [24][25][26][27] ─ prediction of activity delays [28] ─ risk prediction [29,30] ─ cost prediction [31] These studies use different machine learning approaches such as decision trees [21], support vector machines [16], Markov models [12], evolutionary algorithms [14] among others.…”

A Novel Business Process Prediction Model Using a Deep Learning Method

“…The following studies address the next process event prediction that we investigate in this paper. A multi-stage model, which starts by clustering event sequences using the k-mean algorithm combined with sequential alignment, builds individual Markov models on the obtained clusters (Le et al 2014). Experiments were conducted on records of processes obtained from a telecommunication company.…”

“…An approach by Le et al (2017) uses sequential k-nearest neighbor classification and an extension of Markov models to predict the next process steps by considering temporal features. Using the same process log data as Le et al (2014), they showed the superiority of this model over Markov and Hidden Markov Models (HMM). Unuvar et al (2016) proposed a decision tree model to predict the next activity in running instances of processes with parallel execution paths.…”