IBM SPSS Modeler – a wide set of analytics techniques, visualisation of results, and integration with databases: SQL push-back, In-Database Mining, Hadoop, and Spark.
Scoring and other analytical processes can be run by business users.
Storage, publishing, group work, and versioning of objects and processes.
Distribution of results, assessments saved in databases, or recommendations that are available on-line via operating systems.
Analyses, predictive model learning, scoring, advanced scheduling.
Solution architecture that ensures flexible adjustments to organisational requirements and integration of predictive analyses with business processes and systems.
The benefits of running complex operations on huge data volumes by a single person who directly uses results in a quick and easy manner employing feedback is invaluable.
Risk models of PS CLEMENTINE PRO involve key hazard risk domains: new customer risk assessment, customer monitoring, and transaction monitoring. The system can be easily integrated with operating systems of your organisation and adjusted to your business processes without causing any delays related to additional procedures. It complies with state and European regulations.
The high-risk case identification system facilitates defining detection scenarios for various risk domains, including credit, internal, web, card, transaction frauds, etc. in a flexible way. The open architecture of PS CLEMENTINE PRO facilitates detailed calibration of implemented risk assessment scenarios in order to adjust them to the nature and size of identifiable threats. The risk assessment process may be carried out on-line or in batch mode depending on the area.
Customer Relationship Management
By using advanced algorithms of PS CLEMENTINE PRO you can have a huge impact on such areas as income or campaign response increase, limiting customer churn, or acquisition of new accounts. This facilitates achieving business goals by correct determination of analytical goals and utilisation of information contained in data. Results can be made available as a report or predictive model to support off-line and on-line activities.
Debt Collection Management
A single solution that enables the analyst to manage and perform predictive analyses based on historic data using PS CLEMENTINE PRO. It can calculate risk scores automatically which can then be integrated with third-party systems. It has integrated predictive models to assess the chance of debt recovery at the initial stage of the debt collection process. It also has implemented scenarios utilising historic data to search for ineffective elements in action strategies or to create strategy assessment with the champion-challenger approach.
Facilitates programming analytical data flows in a clear graphic environment. Handles the whole data mining process; from accessing data to implementation of results. The entire process in one tool. The program provides easy access to data from various sources (databases, logs, etc.) and facilitates efficient data preparation and processing for modelling. It enables the user to quickly build and evaluate models based on a wide set of statistical and data mining techniques or third-party tools and database-native algorithms. Implementation of results may be carried out through reports, assessments saved in databases, or recommendations that are available on-line via operating systems.
PS CLEMENTINE Database is used to store, organise, and make available system analytical objects. It facilitates efficient management of IBM SPSS Modeler flows and flows secured by authors of models. It also stores schedules, which are definitions of the modes of activation of individual tasks by PS CLEMENTINE PRO and tasks: flows or sets of flows for automation purposes. PS CLEMENTINE Database uses MS SQL Server Express by default, but can be implemented to support virtually every database platform.
PS CLEMENTINE Scheduler is an application responsible for activating PS CLEMENTINE PRO tasks in accordance with defined schedules. This way many regular processes can be automated and provide a constant flow of valid information to decision-makers or feed organisation's operational systems.
Grouping algorithms are used to separate homogeneous subgroups in a population. For example, they can segment customers by the way they use company services. This group of algorithms includes k-means clustering method, two step cluster analysis, Kohonen networks, and anomaly detection.
The role of Rule induction algorithms is to identify which elements occur together. A typical application is market basket analysis which determines which products are bought together by a consumer of goods or services. In PS Clementine these tasks can be performed with such tools as GRI, Apriori, CARMA, and the sequence detection algorithm.
Classification algorithms determine affinity to one group based on a set of characteristics. They can be used to indicate customers who will respond positively to an offer or who are considering resigning from company services, also to assign credit risk, or detect fraud. These algorithms include discriminant analysis , C&RT, CHAID, QUEST, C5.0, decision lists, binary classifier, neural networks, and a self-learning Bayesian response model.
Algorithms used for prediction estimate quantifiable values based on specified characteristics. For example, they can estimate customer value, their age, or income. Prediction can utilise such tools as regression analysis, logistic regression, generalised linear model, time series analysis, and neural networks.
CRISP-DM (CRoss-Industry Standard Process for Data Mining) is a project implementation methodology and data mining process model that provides a complete and universal map for projects aimed to create business solutions using data mining techniques. PS CLEMENTINE PRO provides full support of this model and organises work.
• Uderstanding business - tasks in this step involve understanding business conditions, mapping business goals onto analytical targets, and preparing a project plan.
• Understanding data - the aim of this step is to identify data sources, get familiar with them, and understand their meaning for the business.
• Preparing data - tasks in this step lead to creation of final datasets that conform to format requirements, measurement scale, or cleanliness for modelling, and their description for clear interpretation of models.
• Evaluation - the aim of this step is to verify models in light of business criteria; select the best models for distribution; and make decisions about the next steps and changes in future iterations of the process.
• Distribution of results - – This key stage is to provide recipients with results, which could simply mean the delivery of a report. In data mining projects, however, it more often involves additional information being written in databases or use of models to provide recipients with e.g. on-line business recommendations.
Predictive Solutions (formerly known as SPSS Polska) has been providing solutions for extracting information from data for 25 years. It offers knowledge and experience together with software and comprehensive solutions for efficient data analysis in business, public administration, research, and education.
The solutions use IBM SPSS technologies, whose main distributor in Poland, since 1991, is Predictive Solutions. Dedicated solutions for selected business areas: anti-money laundering (PS AML), customer relationship management (PS ACRM), fraud prevention (PS FRAUD), and others were created based on these technologies. The company also provides original solutions, such as: comprehensive analytic and reporting platform - PS IMAGO PRO and case management system (PS SYMOBIS).
From statistics and reporting, to data mining and predictive solutions, Predictive Solutions helps you use your data to look into the future and make the best decisions.
Predictive Solutions Sp. z o.o. [formerly SPSS Poland]
ul. Racławicka 58 ⋅ 30-017 Kraków, Poland
tel. +48 12 636 45 35 ⋅ fax 102