Trip attraction model of central market in Poso City based on multiple linier regression model

Analysis of trip attraction in shopping centers is one of the most important aspects of travel demand management (TDM) in developing countries. This research aims to describe significant factors that influence people in obtaining their choice of frequency. Variables used in this research are family’s socio-demographic variables, properties of trip to shopping centers, nature of selecting trip time, and ways to travel. This research adopts multinomial logit model by modeling or building causal relationship between significant variables that influence trip frequency. The findings show that on holiday, no variable has significant influence towards the trip attraction of visitor movement at the Poso Central Market, while on workday, shopping cost (X16) has become the most influential variable. Based on the regression equation, the trip attraction of visitor movement model at the Poso Central Market on holiday is Y = 2.076 + 0.997 X6 +0.276 X13 + 0.605 X15 + 0.643 X16, where R2 = 0.004 below 5% or tend to close to 0, so it can be concluded that the ability of independent variables in explaining variation of variables is very limited. Meanwhile, the regression equation on the trip attraction of visitor movement model at the Poso Central Market on workday is Y = 3.090 + 0.250 X2 +0.158 X6 + 0.628 X15 +0.050 X16 +0.662 X17, where R2 = 0.030 below 5% or tend to approach to close to 0, so it can be concluded that the ability of independent variables in in explaining variation of variables is also very limited.


Introduction
Of all variety of human activities, one of them is trip to fulfill the needs of the economy, such as shopping trip.The human needs for shopping is very important to fulfill since this plays an important role in fulfilling human being's daily needs, thus enabling high frequencies of trip to a shopping center.Recently, shopping trip has received a wide focus of researchers' attention.Shopping trip has big proportion for urban trips, especially during peak period.The journey has more individual temporal flexibility than travel journeys and provides more congestion and some types of environmental problems in the downtown zone [1].
Trip needs are fundamental in traffic planning.The key is how much effort is needed to meet the travel demand what method should be used.Among all factors that influence the selection of modes, the dominant factors in working is time factor.The next dominant factor is cost, and then safety, and the last factor is convenient factor [2,3].
The function of travel-time constraint is more dominant than the function of mileage constraint and trip cost in the modeling process of transportation needs.Mode is a means used to travel.Models of mode selection aims to determine proportion of people who will use each mode.Socio-economic condition and pattern of traveler's activities is a major factor affecting traveler's decision.Before predicting transportation demand, it is necessary to determine the factors of traveler's decision, in this case how customers' behavior can be used to predict their decision in choosing transportation service.Decisions made by travel agents also greatly determine quantity, distribution of modes and routes, and time of transportation means in the transportation network as well [4][5][6].
Through the analysis of travel demand management (TDM), it is found that several key aspects have become people's interest to solve problems that arise; such is key aspects in four trip models (i.e.trip awakening, trip attraction, trip frequency, and others).For trip frequency, people who come to shopping places consider many factors for their trip (mall or traditional market), not only for one particular place but also for destination of shopping trip.
From various existing shopping centers in Poso, the researchers chose Poso Central Market as the research object.Formerly, this market was located at the center of Poso City, but then has been located in suburban area of Poso City.This relocation also influences the movement behaviors of travelers.This research is expected to generate model of trip frequency that can be used to determine the route for public transportation of the city.

Desk study
The researchers used questionnaire as technique of data collection.The data collection activities were done in Poso Central Market, where the number of respondents was obtained after a survey was conducted to find out the number of visitors entering the market zone.The sampling procedure was divided into workday and holiday, where the samples for workday were obtained on Tuesday, February 2017 and for holiday were on Saturday, February 18, 2017.

Determining Sample Size
In this research, sample size was determined based on the population size of mode transportation service users.The researchers used Slovin's formula [7], which is as follows: 1 Nd

N n + =
Where: n = number of samples d = error tolerance N = total population Sample size was determined based on the population number, which was the number of visitors on workday and holiday.Error tolerance for sample size was 0.05 or 95% for each population number.Table 1 shows the sample size for this research.

Correlation Analysis
Correlation test was used to test the strength of a relationship between independent and dependent variables, and between independent variables.

Regression Analysis
The largest number was chosen to represent all independent variables with correlation value greater than 0.5.Later, all chosen variables were used to yield equation model.

Visitors Distribution
(1) Based on the survey of market visitors conducted on workday and holiday, the distribution of market visitor frequency in time period of 15 minutes with 10 hours of time intensity can be seen in Figure 1 and 2.

Model of trip attraction calculation
The data obtained from questionnaire were analyzed by multiple regression through SPSS 16 software.The model of trip attraction formulation was later generated by regression analysis with the following steps:

Correlation analysis for variables
After performing normality test for each variable, correlation analysis was conducted to analyze the correlation between all independent variables (X) and correlation between dependent variable (Y) with independent variables (X), where the dependent variable (Y) was the intensity of shopping day.Before reading the results of each variable's effect, we have to make sure whether those variables had significant effect on Y or not, by checking the t value and significance value of each variable.If we use t value, it is said to be significant if t value is greater than t table, while if we use significance value, it is said that the variable is significant if sig value is lesser than a (in this case, the value = 5%).We can see in the table that no variables had significant effect on dependent variable.while if we use significance value, it is said that the variable is significant if sig value is lesser than a (in this case, the a value = 5%).We can see in the table that variable X16 had significant effect on dependent variable.Figure 4 shows the graph model for normality test performed with SPSS, where the existing points spread around diagonal line despite the extreme point, and its distribution follows the direction of diagonal line.Based on the independent variables, this regression model can then be used to predict the amount of trip attraction to the market on holiday.

Conclusions
Based on the analysis results, no factors had significant effect on the trip attraction of Poso Central Market visitors on workday, while shopping expense (X16) had significant effect on the trip attraction on holiday.From the regression equation, model of trip attraction for visitor movement to Poso Central Market on holiday was Y = 2.076 + 0.997 X6 +0.276 X13 + 0.605 X15 + 0.643 X16, where R² = 0.004 below 5% or tends to close to 0, therefore it is concluded that the ability of independent variables in explaining variable variation is very limited.Meanwhile, based on the regression equation, trip attraction for visitor movement to Poso Central Market on workday was Y = 3.090 + 0.250 X2 +0.158 X6 + 0.628 X15 +0.050 X16 +0.662 X17, where R² = 0.030 below 5% or tend to close to 0. Therefore, it can be concluded that the ability of independent variables in explaining variable variation is very limited as well.

Fig 3 .Figure 3
Fig 3. Graph of Normality Test for HolidayFigure3shows the graph model for normality test performed with SPSS, where the existing points spread around diagonal line despite the extreme point, and its distribution follows the direction of diagonal line.Based on the independent variables, this regression model can then be used to predict the amount of trip attraction to the market on workday.

Fig 4 .
Fig 4. Graph of Normality Test for Workday

Table 1 .
Sample Size

Data Display and Analysis Method
Several statistical tests were used to analyze the data, as explained below:2.2.1 Validity and Reliability TestQuestionnaire validity and reliability test were used to ensure that questionnaires used in this research can really measure research variables.An instrument is considered as valid if it can measure what is supposed to measure and reveal data from the examined variables.

Table 2 .
Result of Validity and Reliability Test for HolidayR table value can be seen in R table, where df = number of sample -2, or in this case df = 381-2 = 379.From significance level 5%, R table value = 0,10115.The results obtained after analyzing samples for workday showed that variable X2, X6, X13, X15, X16, and X17 were reliable and valid.

Table 3 .
Result of Validity and Reliability Test for Workday

Table 4 .
Matrix of Correlation between Variables on Holiday

Table 7 .
Results of Modeling Visitor Movement Attraction on Workday Similar to the previous procedure, we have to make sure whether those variables had significant effect on Y or not, by checking the t value and significance value of each variable.If we use t value, it is said to be significant if t value is greater than t table,