Data Mining

Service Introduction

We assist businesses to tap client potentials and improve their competitiveness with its gigantic database generated through the years of industrial expertise and by using professional technologies. We also provide consulting services with abundant database resources.

ACMR Services

Client Analysis	Client consumption analysis and forecast Client segmentation and forecast Client value analysis and forecast Client loyalty analysis Client loss analysis and forecast Inter-selling analysis Fraud early-warning analysis
Sales Forecast Analysis	Sales trend forecast Stock balance forecast
Risk Management	Risk credit grading Risk analysis and forecast
Operation Performance Management	Balance scoring card

Target Customers

Banking, Insurance, Telecom, Retail, Manufacturing, Medicine and Energy sectors, as well as government institutions

Service

ACMR data mining operation procedures are based on the Cross-Industry Standard Process for Data Mining (CRISP-DM process), comprising the following steps:

Data Mining – Solution to Business Problems

After fully understand the client's request, we transfer the various requirements into data mining questions and set up a preliminary execution plan. We maintain close contact with the client to set up the data mining targets.

Confirmation of client's request and problems: understand its background, problems and request, and confirm its development objective;
Confirmation of the guidelines for data mining: data mining objectives and criteria for success;
Draw up the project execution plan, evaluate the tools and technologies.

Data Preparation

Since requested data might be incomplete, we must prepare before data mining, which covers fully understand the data specifics, processing (cleaning), transforming, integrate and filtrate the data.

Understanding the Data Specifics

Although companies provide a good deal of data, we need to understand the data specifics and definitions with the help of IT technicians. Meanwhile, we must evaluate whether the supplied data can fully satisfy the client's problems.

Data Processing

Missing value processing: Adjust and process empty numerical values, inexistent numerical values, and abridged data.
Data consistency processing: Sometimes, different field values represent identical meanings, while same fields have different implications. Therefore, we must carry out consistent processing to avoid confusion in data definition.
Data validity processing: Settle problems out of line with valid field values, possibly wrong keyboard input or programs.

Data Transform: Adjust skipped or wrong data according to the data specifics and transform them into the required format.

Establish and Evaluate Data Mining Models

We offer solutions for clients' problems by processing and analyzing data, applying various analytical methods and technologies. We evaluate the models before using them, and re-examine the execution steps to make sure we achieve the clients' targets.

Analyze and Validate Data Mining Results

We assist professionals to analyze and interpret the results and raise analysis proposals to the clients. We submit the mined data as statements and models.