Implementation Of Apriori Data Mining Algorithm to Increase Sales Of Caringin Shop

: The caringin shop is a shop that sells daily household and kitchen needs. Caringin shop sales tend to stagnate because in the last few months there has been no significant increase in turnover and tight business competition where the large number of similar shops has become a problem. This requires a precise and careful marketing strategy. This research implements an a priori algorithm as a marketing strategy. The data used was 146 transactions for 1 week. The analysis was carried out manually using Microsoft Excel and libraries in the Python programming language that supports apriori algorithms packaged in a simple web-based application. The analysis resulted in a conclusion that if a consumer buys sugar, then the tendency is to buy coffee with a confidence level of 65% supported by a support value of 10% of all transactions with a lift ratio of 3.9, if buying coffee, then the tendency is to buy sugar with a confidence level of 62% supported by a support value of 10% of all transactions with a lift ratio of 3.9, and if buying eggs, then the tendency is to buy instant noodles with a confidence level of 61% supported by a support value of 11% of all transactions with a lift ratio of 2.8.


Introduction
The caringin shop is a shop that sells various household and kitchen equipment for daily needs (Setiani et al., 2021).The caringin shop is in South Bengkulu Regency.The caringin shop is managed independently by the shop owner.Caringin stores are small scale stores compared to minimarket franchise stores such as Indomaret and Alfamart.
Technological advances and the proliferation of minimarket franchise stores are the starting point for obstacles to the development of caringin stores (Nisa, Siska Fitriyanti and Dewi Siska, 2021;Novriady and Nasrudin, 2021).
The caringin store experienced several problems that hampered the store's development (Kusak et al., 2021).The problems faced are the decline in shop turnover in the last few months, the number of shop visitors is starting to decline, and tight business competition and a number of similar shops which are not far from the caringin shop (Adebayo & Aziz, 2019).Plus, the proliferation of large-scale store franchises has made https://penerbitadm.pubmedia.id/index.php/KOMITEKcompetition increasingly fierce and small-scale stores such as caringin stores are starting to be abandoned.Based on the three problems above, caringin stores need appropriate and accurate sales strategies so that they can increase store sales (Zhao et al., 2021).
The solution offered in this research is implementing an a priori algorithm.The a priori algorithm produces an analysis of the relationship between products and other products and predicts product sales patterns (Br Sembiring and Sembiring, 2023).The results of this research are in the form of additional knowledge for store managers in developing business strategies in the form of sales links between products, product layout, and recommendations for product discount packages (Zhou, 2020).
The aim of this research is to produce a sales strategy for caringin stores through the implementation of an a priori algorithm.This aims to increase sales turnover, provide a better experience to customers (Xie, 2021), recommend products that are right for customer needs, and provide insight and basic considerations in developing stores based on accurate data (Li et al., 2021).

Problem Analysis
Researchers have conducted in-depth interviews with the owner of the Caringin Shop and produced several problem points that were captured (Verma et al., 2020).These points are that there is a tendency to decrease turnover but it is not significant.This is due to the mushrooming of similar shop franchises which are starting to appear around the caringin shop area.In addition, caringin store owners have minimal knowledge regarding store sales strategies based on advanced research and analysis (Sadikin, 2024).Based on information from the shop owner, the researcher is interested in raising the research topic of increasing sales by implementing the a priori algorithm (Abdullah et al., 2019).

Implementation Method
This research uses various implementation methods.The implementation methods used in this research are as follows: 1. Problem Analysis Researchers began to identify the root of the problem in the research theme, namely increasing product sales.As a result of identifying the root of the problem in this research, the researcher confirmed the findings with the shop owner through in-depth interviews.Apart from that, researchers propose solutions that become research proposals.

Data Collection
At this stage, researchers collected data through interviews, surveys and literature studies.Primary data was obtained from interviews and surveys at caringin stores.Secondary data was obtained through literature studies by studying data and information from various information sources related to data mining and a priori algorithms.Apart from that, researchers also determine the population and data samples that will be used in the data analysis stage (Saxena & Rajpoot, 2021).

Data Processing
At this stage, researchers carry out data selection, data preprocessing, and data transformation.

Data Analysis
At this stage, researchers begin to carry out data mining analysis using a priori algorithms through the stages of high frequency pattern analysis and associative rule formation.

Evaluation and Drawing Conclusions
At this stage, researchers carry out evaluations using the lift ratio and draw conclusions based on the results of data mining analysis.

Documentation
Researchers in the documentation stage prepare research reports in order to complete the entire series of research (Jhang et al., 2019).

System Planning
At this stage the researcher designs a web-based application that contains the implementation of an a priori algorithm.The system designs developed are flowcharts and mock ups.

Flow Chart
Before designing the application, the researcher created a flow chart to make it easier to implement the a priori algorithm.The flow chart of the implementation of the a priori algorithm is shown in Figure 3.1.

Implementation
The implementation stage contains various sub-stages that form the entire implementation of the a priori algorithm.The stages that are passed include the following.

Data Understanding
Researchers carry out data understanding to find out and understand what data is needed in analysis using a priori algorithms to achieve business goals and obtain better sales strategies.The data required is sales transaction data per product which includes the attributes transaction/invoice number, product name, transaction date.The transaction/invoice number shows the unique ID or primary key of the sales data.The product name represents the product being transacted.Transaction date shows data on the date when the transaction occurred.

Data Processing
Researchers at the data processing stage carry out data selection, data cleaning, determine data attributes according to the stages of data understanding, and carry out data transformation.Researchers also eliminate data if they do not have complete data attributes, for example, invoice numbers are not available, dates are not listed, and product names are empty.The data for 146 transactions obtained and taken during the 1 week period from April 6 -12 2024 is shown in Table 4.1.

Conclusion
A priori algorithms can provide alternatives for increasing sales based on association strategies with historical data on products purchased by customers.This research implements an a priori algorithm wrapped in a web-based application, making it easier for shop owners to carry out analysis to obtain alternative sales strategies.The a priori algorithm has succeeded in providing knowledge or insight to shop owners that there is a relationship between one product and another so that shop owners can make decisions based on the analyzed data.