文档库 最新最全的文档下载
当前位置:文档库 › Incremental Update on Sequential Patterns in Large Databases

Incremental Update on Sequential Patterns in Large Databases

Incremental Update on Sequential Patterns in Large Databases
Incremental Update on Sequential Patterns in Large Databases

Incremental Update on Sequential Patterns in Large Databases

Ming-Yen Lin and Suh-Yin Lee

Institute of Computer Science and Information Engineering

National Chiao-Tung University

Abstract

Mining of sequential patterns in a transactional database is time-consuming due to its complexity. While maintaining present patterns is a non-trivial task after database update, since appended data sequences may invalidate old patterns and create new ones. In contrast to re-mining, the incremental update algorithm proposed which effectively utilizes discovered knowledge is the key to improve mining performance. By counting over appended data sequences instead of entire updated database in most cases, fast filtering patterns found in last mining and successive candidate sequence reductions together make efficient update on sequential patterns possible.

1.Introduction

As computerized applications facilitate automatic collection of large amount of data into databases, new challenges to extract practical knowledge fro m large databases arise. Data Mining and Knowledge Discovery in Databases (KDD) thus has been recognized as a promising field of AI, statistics and database researches [1].

Mining of sequential patterns was first introduced in [2]. The purpose is to discover sequential patterns in a database of customer transactions, consisting of records having customer id, transaction time, and transaction items.

A sequential pattern indicates a sequence of transactions that usually occurred serially in time. For example, in a transactional database, one might purchase a PC and then purchase a printer at some later time. After a period of time, he/she could possibly buy some printing software and a scanner. If there exists a sufficient number of customers in the database who have a purchasing sequence of PC, printer, printing software and scanner, then such a sequence is a sequential pattern. Note that items in a sequential pattern need not be simple items, and that transactions need not be consecutive. As pointed out in [11], this problem was motivated by retailing applications, but the results can be applied to many business and scientific domains.

In general, the process for sequential pattern mining is comprised of several passes. Before mining process starts, the user specifies the minimum support, where the support is the percentage of data sequences that contain the pattern. Note that the support in association rules is transaction-based, while in sequential patterns, the support is based on data sequences. The counting of support is different. Suppose that a customer has two transactions that bought the same item. In the discovery of association rules, the customer contributes to the support count of that item by two, whereas it is counted only once for the support count in sequential pattern mining [2].

The transaction database is sorted by customer id into customer sequences, or data sequences. The mining process makes several passes over these data sequences. Various candidate sequences, which are potential sequential patterns, are generated in individual passes. In each pass, all data sequences, or trimmed data sequences, are scanned to increment the supports of contained candidate sequences. Candidate sequences having

minimum support become frequent sequences. They are then used for the creation of longer candidate sequences for next pass. The algorithm stops when there is no candidate sequence any more. The performance of mining depends heavily on the number of candidate sequences and the number of data sequences in a pass.

The issue of maintaining sequential patterns becomes essential because database transactions may be updated over time. Due to new transactions, some existing sequential patterns would become invalid after database update since they might no longer have sufficient supports, while some new sequential patterns might appear. However, not much work on incremental updating of sequential patterns can be seen. In order to ascertain sequential patterns up to date for the updated database, re-execution of m ining algorithm on the updated database is required. Nevertheless, because of appended database transactions, re-execution of mining algorithm demands more time than previous mining. Moreover, the effort of mining last time is wasted if all discovered sequential patterns in the original database were ignored.

There are many excellent algorithms that deal with the mining of association rules [1, 4, 10, 12] and several algorithms were developed for the mining of sequential patterns [2, 11, 13]. Algorithms to discover frequent episodes in a single long sequence and its generalization can be found in [8, 9]. There is no efficient algorithm designed for the maintenance of sequential patterns in large databases. As mentioned earlier, they call for the need of re-mining the whole database. Incremental updating techniques for the maintenance of association rules were proposed in [5, 7, 12]. However, previous work dealing with the incremental updating of sequential patterns cannot be found. Besides, appended transactions induce more complicated problems in sequential pattern mining than in association rule mining. The problem of finding association rules concerns with intra-transaction patterns whereas that of sequential pattern mining concerns with inter-transaction patterns [2]. Appended transactions bear no relation to original database for the former problem, while for the latter problem, different transactions with same customer id in both databases must be sorted into one data sequence.

In this paper, we propose a new algorithm that can utilize information about discovered sequential patterns, and efficiently reduce candidate sequences for mining in updated database. The objective of this work is to solve the update problem of sequential patterns after a nontrivial number of new transactions have been appended to original database. Assuming that minimum support keeps the same, existing frequent sequences and their supports in original database could be utilized for the mining of updated database. Through effective reuse of previous derived knowledge in each pass, the number of candidate sequences is substantially reduced. Instead of counting all candidate sequences for full updated database, counting reduced candidate sequences for original database and for increment database, could achieve better performance. For transactions that have same customer id in both original database and increment database, we extract old transactions from original database and merge into increment database. Since the size of increment database is smaller than original database in general, better performance thus could be retained.

The remaining paper is organized as follows. Section 2 describes the problems of sequential pattern mining and incremental update. In Section 3, we propose our n ew algorithm for the maintenance of sequential patterns. Section 4 concludes our study.

2.Problem Formulation

2.1 Sequential pattern mining

Let database D be a set of transactions, where each transaction T consists of transaction-time (tid), customer-id (cid), and a set of items. The quantities of items are not considered here. For simplicity, we assume that no customer has more than one transaction at same transaction time.

An itemset I, denoted by (x1, x2,…, x m), is a nonempty set of items, where each item x k is represented by an integer. A sequence s, denoted by , is an ordered set of itemsets. Each a j is an itemset, which is called an element of the sequence. An item can occur only once in an element of a sequence, but can occur multiple times in different elements of a sequence. A k-sequence consists of k items. For example, <(8)(2)(1)> and <(3)(5,9)> are both 3-sequences.

A sequence is a subsequence of another sequence if there exist integers i1

1

, a2?b i

2

,…, a n?b i

n

. For example, <(2) (5,9) (8)> is a subsequence of <(6) (2) (1) (5,9) (3,8)> since (2) ? (2), (5, 9) ? (5, 9) and (8) ? (3, 8). However, <(3) (8)> is not a subsequence of <(3,8)>, and vice versa.

All transactions from the same custom er in the database can be grouped together, and then sorted by transaction time in increasing order into a customer sequence, or called data sequence. The support for a sequence s is defined as the fraction of total data sequences that contain s. A data sequence contains a sequence s if s is a subsequence of the data sequence. A sequence is a frequent sequence if its support is greater than the user-specified minimum support, denoted min_sup.

Given a database of customer transactions, the

problem of sequential pattern mining is to find all frequent

sequences. Concerning association rules, the discovery of sequential patterns can be thought of as association discovery over a temporal database.

2.2 Incremental update on discovered sequential patterns

Let |DB| be the number of data sequences in the original database DB and min_sup be the minimum support. After some update of the database, a few transactions are appended to DB. These transactions can be sorted by cid into |d b| data sequences in the increment database db. UD is the updated database combining all data sequences from DB and db, UD = DB∪ db. Let there be |b| customers appearing both in DB and db. If all customers in db are new customers with respect to DB, that is, all cid s in db are different from those in DB, |b| is zero. Let S k DB be the set of all frequent k-sequences in DB, S k UD be the set of all frequent k-sequences in UD, and the set of all sequential patterns in DB and UD be S DB and S UD respectively.

Assume that for each sequential pattern s in DB, its support count, denoted by S count DB, is available. With respect to the same minimum support min_sup, a sequence s is a frequent sequence in the updated database UD, if its support count S count UD is greater than min_sup× (total data sequences in UD). That is, S count UD is no less than min_sup× (|DB|+|db|-|b|). X count DB, X count db and X count UD are support counts of a sequence X in DB, db and UD. If there is no same cid in DB and db, X count UD = X count DB + X count db. A sequential pattern s in S DB might not be in S UD because of database update. On the other hand, a sequence s’ that is not in S DB might turn out to be in S UD.

Consequently, the problem of incremental update on sequential patterns is to find the new set S UD of frequent sequences in UD. As shown in Fig. 1, In order to find S UD, previous approaches take all data sequences to compute sequential patterns and their supports. With discovered S DB and support counts, the incremental algorithm updates supports of sequential p atterns in S UD by scanning data sequences in db only, if they were frequent in DB. New candidate sequences, which might not exist in DB, are generated in the same scan. Some new candidate sequences are pruned, if they do not have sufficient supports relative to db, before the scan of DB starts. The complexity of mining process is reduced due to the employment of previous knowledge and the increment database.

Updated database, UD = DB + db

data sequence 1

data sequence 2

DB .

.

.

data sequence

db data sequence

UD

S U D?

...

Original database, DB

data sequence 1

data sequence 2

DB .

.

.

data sequence|

+

Increment database,

data sequence 1

db data sequence 2

data sequence |db|

DB

(b) incremental updating

Figure 1: Incremental update versus re-mining 2.3 Example of sequential patterns after database updated

Consider a database with transactions sorted by customer id into 6 data sequences as shown in Fig. 2. Assume minimum support is set to 33%, i.e., minimum support count of data sequences being 2. The sequential patterns are <(1)>, <(2)>, <(3)>, <(4)>, <(5)>, <(8)>, <(1,2)>, <(8)(2)>, <(8)(1)>, <(2)(1)>, <(2)(8)>, <(5)(1)>, <(5)(3)>, and <(3)(1)>.

After some update activities, assume the transactions appended to the database all come from new customers. The increment database, sorted by cid into data sequences, is shown in Fig. 3. With same minimum support, it requires 3 data sequences to be a frequent sequence now. Previous sequential patterns <(3)>, <(8)(2)>, <(2)(1)>, <(2)(8)>, <(5)(1)>, <(5)(3)>, and <(3)(1)> are no longer frequent due to this update. While <(7)> and <(8)(7)> become new sequential patterns because they have

minimum supports now.

Customer Id Data Sequence

1 <(8)(2)(1)>

2 <(1,2)(8)(7)>

3 <(1,4)>

4 <(1)(4)>

5 <(5)(3)(1)>

6 <(2,5)(8)(3)(1,2)>

Figure 2: The original database DB with 6 data sequences

Customer Id Data Sequence

7 <(5,7,8)(7)>

8 <(8)(1)(7)>

9 <(1,2)(4)>

Figure 3: Data sequences in the increment database db

In cases of update when the customer id of new transactions appears in the original database, these transactions must be appended to the same customer to form his/her data sequence. For example, assume the customers, whose cid s are 1 and 4, bought item 7 at later time and the data sequences for cid=1 and cid=4 now are <(8)(2)(1)(7)> and <(1)(4)(7)> respectively. Fig. 4 shows an example of data sequences, which come from both old and new customers, in the increment database. The data sequences for mining in updated database UD is shown in Fig. 5.

After invalidating sequences <(5)> and <(8)(1)>, the resulting se quential patterns in this database are <(1)>, <(2)>, <(4)>, <(7)>, <(8)>, <(1,2)> and <(8)(7)>, for the given minimum support 33%.

Customer Id Data Sequence

1 <(7)>

4 <(7)>

10 <(4)>

11 <(1,2)>

12 <(8)(7)>

Figure 4: Data sequences in db with cid s occurring in DB

Customer Id Data Sequence

1 <(8)(2)(1)(7)>

2 <(1,2)(8)(7)>

3 <(1,4)>

4 <(1)(4)(7)>

5 <(5)(3)(1)>

6 <(2,5)(8)(3)(1,2)>

7 <(5,7,8)(7)>

8 <(8)(1)(7)>

9 <(1,2)(4)>

10 <(4)>

11 <(1,2)>

12 <(8)(7)>

Figure 5: Merged data sequences in UD

3.Fast Sequential Pattern Update Algorithm

(FASTUP)

3.1 Previous algorithms versus FASTUP algorithm

Apriori algorithm was designed to discover association rules [1], while AprioriAll algorithm introduced in [2] was the first algorithm that deals with the problem of sequential patterns mining. In subsequent work [11], the same authors proposed GSP(G eneralized S equential P attern)algorithm that outperforms AprioriAll. An algorithm called SPADE (S equential PA ttern D iscovery using E quivalence classes) was designed that uses simple

join operations to find sequential patterns [13]. However, vertical database layout was used, instead of horizontal database layout used in earlier approaches [2, 11]. Therefore, the GSP algorithm is briefly reviewed here.

GSP algorithm makes multiple passes over the database. In the first pass, frequent 1-sequences with their supports are determined by counting the support of 1-itemsets. In the subsequent passes, candidate k-sequences are generated from frequent (k-1)-sequences obtained in pass-(k-1). Then the supports for these candidates are computed and those with minimum support become frequent sequences. This process is iterated until no more candidate sequences are formed. In each pass, every data sequence is checked to increment the support count of candidates contained in this data sequence. Hence, there are two essential sub-processes in this algorithm:

1.Candidate generation: Let L k denote

the set of all frequent k-sequences, and C k

denote the set of candidate k-sequences. Given

L k-1, C k is generated by selectively join L k-1 with

L k-1 itself. If there exist any (k-1)-subsequence

of a candidate which is not in L k-1, the candidate

is pruned from C k. Candidates are inserted into a

hash tree to enable fast counting. Please refer to

[11] for the detailed join operation and the hash

tree mechanism.

2.Support counting: Every item and its

following items in the checking data sequence is

hashed to reach leaf buckets in the candidate

hash tree. The support of each candidate in the

buckets, if the candidate is contained in the data

sequence, is incremented.

Note that as the name GSP suggests, it not only solves problems formulated in previous section, but it also solves problems generalized with constraints on pattern hierarchy and transaction-time [11]. The algorithm described here is to solve the fundamental problem of sequential patterns without constraints.

The basic construct of our algorithm FASTUP is similar to that of GSP, with imp rovements on candidate generation and support counting. FASTUP algorithm has several arguments similar to the incremental association update algorithm FUP [5]. Nevertheless, the effect of candidate reduction is more dramatic in sequence mining. Moreover, w e have to consider whether appended data sequences are required to merge with old data sequences. Features that distinguish FASTUP from GSP are listed as follows.

1.During pass-k, for each X∈S k DB, The support

count of X is updated against db, w ithout re-scan

DB. Old frequent sequences that do not satisfy

new support count are filtered out.

2.During pass-k, if X∈S k-1DB and X?S k-1UD, prune

every candidate C∈C k such that X is a

subsequence of C. With this feature FASTUP has

greater capability in candidate reduction by a

simple check in db before counting starts in DB.

3.During pass-k, a k-sequence X, where X?S k DB,

with X count db, is added to C k if

X count db≥ min_sup×|db|. For each X∈C k, The

support count of X is updated against DB. The

generation of candidate sequence X requires less

data sequences checking in db than in DB or in

UD. Only the supports of “new” candidates,

instead of all candidates, need to be checked in

DB.

These features altogether enable FASTUP to have better performance than re-execution GSP on updated database. Previous frequent sequences and candidate sequences in DB that do not belong to UD are pruned

away by simple checks on db. Without missing any potential new frequent sequences in UD, fewer new candidates are generated from db for checking against DB.

3.2Merge data sequences of the same customer for sequential pattern mining

For the given data sequences in DB and db, the sequential patterns in UD can be found by re-execution of GSP. Transactions that come from the same customer, either in DB or in db, are parts of the unique data sequence for that customer in UD. Data sequences with same customer id in both databas es must be merged into one data sequence before GSP starts. Likewise, FASTUP algorithm merges data sequences when necessary but relays practical information about discovered sequential patterns.

In order to retain as much information as possible, and to keep smaller number of data sequences for later mining process, we extract data sequences from DB and merge them to db. The merging is accomplished as follows.

1.With given data sequences, sorted by cid, in db

and DB, we check cid s in db against DB to find

any existence of that cid.

2.For the found data sequence ds in DB, the

support counts of all frequent sequences

contained in ds are decremented by one.

3.The ds is merged into that data sequence ds’

with same cid in db, with ds followed by ds’.

Apparently, with sorted data sequence, searching cid could be done in very short time. With known sequences for checking, decrement operations add slightly overhead in merging process. After extraction and merging, there is no data sequence in DB and db with same customer id. FASTUP algorithm then could proceed to next stage.

3.3FASTUP algorithm

3.3.1 Pass-1: find frequent 1-sequences in updated database.

1.Scan db for all 1-sequence X to get X count db.

2.If X∈S1DB∧X count DB+X count db< min_sup×(|DB|+|db|), put

X to the failed set, otherwise add X to S1UD.

3.In the same scan, if X?S1DB∧X count db< min_sup×|db|, X

could not be frequent; otherwise, add X to C1.

4.Scan DB for each X∈C1 to get X count DB. If X count DB

+X count db≥ min_sup×(|DB|+|db|) at the end of pass-1, add X to S1UD.

In comparison with GSP, FASTUP discovers previous frequent 1-sequences that is still frequent, filters out those that are invalid now, and generates potential candidate 1-sequences within a scan on increment database. Every new candidate in the set of new candidate 1-sequences is then checked against the original database to see if it is frequent. Contrast to GSP, GSP takes every item as a candidate and counts over the whole database. FASTUP is obviously faster than GSP in this pass.

3.3.2 Pass-k: find frequent k-sequences in updated database.

1.Generate C k from S k-1UD as GSP described [11]. C k is

characterized into two subsets. Let C k1= C k∩S k DB, C k2=

C k?C k1. C k1 consists of candidates that are also frequent

k-sequence in DB. While C k2 consists of remaining candidates that are “newly” generated.

2. A previous frequent k-sequence X in DB, which

X∈(S k DB?C k1), is not contained in C k. Such X, though in

S k DB, need not be checked against db.

3.Scan db for each X in C k to get X count db. For each X in

C k1, since X count DB is available without checking DB, X

is added to S k UD if it has minimum support.

4.In the same scan of db, for each X in C k2, if

X count db< min_sup×|db|, X could not be frequent. Such X is rem oved from C k2. This step reduces the number of candidates to be checked against DB.

5.Scan DB for each X∈C k2, compute X count DB. X is added

to S k UD if X count DB +X count db≥ min_sup×(|DB|+|db|) at the end of pass-k.

The above process is iterated until no more candidates are generated. At each pass, FASTUP updates supports of frequent sequences in DB which are still frequent in UD, and generates most likely frequent candidates within one scan over small increment db. Candidate sequences, which do not have sufficient supports relative to db, are pruned before they are verified by DB. The set of candidates that FASTUP generated for further checking with DB is smaller than GSP. Consequently, FASTUP could be much faster than previous algorithms of sequential pattern mining.

3.3.3 Complete FASTUP algorithm.

Algorithm Merge_Data_Sequences.

/* Let current data sequence be ds db */

/* Let cid of current data sequence be cid */

For_all data sequences in db do

If cid is found in data sequence ds DB in DB then

For_all frequent sequences contained in ds DB do

Decrement support count of this frequent sequence by 1

End_for

ds db is replaced by ds DB merged with ds db, where transactions in ds DB followed by

those transactions in ds db

End_if

End_for

Algorithm_Fast_Update_Sequential_Patterns.

/* PASS-1 */ /* Initialize all X count db to zero, X is 1-sequence, i.e.,

1-itemsets. */

For_all data sequences in db do

Increment X count db by 1 if current data sequence contains X

End_for

For_all 1-sequence X do

If X is frequent in DB, then

If X count DB + X count db≥ min_sup×(|DB|+|db|), add X to S1UD

End_if

If X is not frequent in DB, then

If X count db≥ min_sup×|db|, add X to 1C

End_if

End_for

For_all X in

1

C do

Initialize X count DB to zero

End_for

For_all data sequences in DB do

For_all X in

1

C do

Increment X count DB by 1 if current data sequence contains X

End_for

End_for

For_all X in C1do

If X count DB + X count db≥ min_sup×(|DB|+|db|), add X to S1UD

End_for

/* All frequent 1-sequences and their counts accumulated */

/* Pass-2 and beyond */

k = 2

Generate C k from S k-1UD

While C k is not empty do

/* Initialize all X count db to zero, X is k-sequence. in C k */

For_all data sequences in db do

Increment X count db by 1 if current data sequence contains X

End_for

For_all k-sequence X in C k do

If X is frequent in DB, then

If X count DB +

X count db≥ min_sup×(|DB|+|db|), add X to S k UD

End_if

If X is not frequent in DB, then

If X count db≥ min_sup×|db|, add X to C k2

End_if

End_for

For_all X in C k2do

Initialize X count DB to zero

End_for

For_all data sequences in DB do

For_all X in C k2do

Increment X count DB by 1 if current data sequence contains X

End_for

End_for

For_all X in C k2 do

If X count DB +X count db≥ min_sup×(|DB|+|db|), add X to S k UD

End_for

/* All frequent k-sequences and their counts accumulated */

k = k + 1

Generate C k from S k-1UD

End_while

3.3.4 Examples on merging and FASTUP algorithm. Example 1. Given a original database DB which combines sequences in Fig. 2 and Fig. 3, with total 9 data sequences. If an increment database db in Fig. 4 with total 5 sequences is appended, the sequences of cid=1 and cid=4 would be extracted from DB and merged into db. Now DB has 7 and db has 5 data sequences. The count for <(8)>, <(2)>, <(1)>, and <(8)(1)> is decremented by 1 due to the extraction of data sequence with cid=1. Likewise, extraction of data sequence with cid=4 would cause support of <(1)> and <(4)> decreasing. After extraction and merging, FASTUP can be applied correctly since no cid appears in both DB and db.

Example 2. Given a database DB, and an increment database db, such that after data sequences are merged, |DB|=1000 and |db|=125. For min_sup=8%, in last discovery, after extraction and merging, found sequential patterns are listed in Table 1.

Let us consider sequential patterns related to items 1, 2, 3, 4, 5. In order to find frequent 1-sequences in UD, a scan on db is made at first. Table 2 shows the result of the scan. Now <(1)> and <(3)> remain frequent with counts updated to 120 and 95, both counts > 8%×(1000+125), i.e.

90. <(2)> is no longer frequent since its count is only 85. <(5)> cannot be frequent because its count in db, being 9, is less than 8%×125. A scan on DB for new candidate

<(4)> is then conducted. Suppose <(4)>count DB=70, <(4)> is a new frequent 1-sequence with count 94. Before proceeding to 2-sequence counting, <(1,2)>, <(1)(2)> and <(3)(2)> are eliminated from counting in db for having subsequence <(2)>, which is not frequent in UD. In fact, candidate 2-sequences now consists of old frequent 2-sequences <(1,3)> and <(1)(3)>, new candidate 2-sequences <(1,4)>, <(1)(4)>, <(4)(1)>, <(3,4)>, <(3)(4)>, <(4)(3)>, and <(3)(1)>. Note that the support count for <(3)(1)> is not available since it failed to be frequent in DB and its count was not kept. After a scan on db for these candidates, the result is shown in Table 3. Now <(1)(3)> remains frequent with count=93. Another counting on DB is only required for <(1)(4)> and <(3)(4)> to check if they have minimum support. It is not necessary to check <(3)(1)> since its count in db is less than minimum required. The discovery continues until no more candidate sequences are generated.

Frequent sequences Sequence Support Count

<(1)> <(1)> count DB=100

<(2)> <(2)>count DB=83

S1DB

<(3)> <(3)> count DB=87

<(1,2)> <(1,2)> count DB=82

<(1,3)> <(1,3)> count DB=81

<(1)(2)> <(1)(2)> count DB=83

<(1)(3)> <(1)(3)> count DB=86

S2DB

<(3)(2)> <(3)(2)> count DB=81 Table 1. Sequential patterns and their counts in DB after merging and extraction

Sequence Support Count

<(1)> <(1)>count db=20

<(2)> <(2)>count db=2

<(3)> <(3)>count db=8

<(4)> <(4)>count db=24

<(5)> <(5)>count db=9

Table 2. 1-sequences and their counts in db

Sequence Support Count

<(1,3)> <(1,3)> count db=2

<(1)(3)> <(1)(3)> count db=7

<(3)(1)> <(3)(1)> count db=8

<(1,4)> <(1,4)> count db=5

<(1)(4)> <(1)(4)> count db=11

<(4)(1)> <(4)(1)> count db=9

<(3,4)> <(3,4)> count db=6

<(3)(4)> <(3)(4)> count db=22

<(4)(3)> <(4)(3)>count db=3

Table 3. Candidate 2-sequences and their counts in db

4.Conclusion and Future Works

Due to the nature of sequence permutation, the problem of sequential pattern mining is more complicated than the discovery of association rules. Without maintenance, validity of discovered patterns may change after update on database. W e propose FASTUP algorithm that efficiently solves the problem by incremental updating without re-mining the whole updated database from scratch. Using frequent sequences and support counts discovered from original database, FASTUP rapidly updates frequent sequences and their counts by scanning over increment database instead of whole updated database. Fewer but more promising candidates are generated by just checking counts in increment database.

The simulation of our algorithm on synthetic data is underway. Given the analysis on sequential pattern mining, the performance of the algorithm could be much faster than previous algorithms for the maintenance of sequential patterns. Further researches could be extended to problems of various minimum supports and problems of generalized sequential patterns such as patterns with is-a hierarchy.

References

[1] R. Agrawal, H. Mannila, R. Srikant, H. Toivonen and A. I.

Verkamo, Fast Discovery of Association Rules, Advances in Knowledge Discovery and Data Mining, edited by U. M.

Fayyad et al, AAAI/MIT Press, pp. 307-328, 1996.

[2] R. Agrawal and R. Srikant, Mining Sequential Patterns,

Proceedings of the 11th International Conference on Data Engineering (ICDE’95), pp. 3-14, Taipei, Taiwan, 1995. [3] R. Agrawal and J. Sh afer, Parallel Mining of Association

Rules, IEEE Transactions on Knowledge and Data Engineering, Vol. 8, No. 6, pp. 962-969, Dec. 1996.

[4] S. Brin, R. Motwani, J. Ullman and S. Tsur, Dynamic Itemset

Counting and Implication Rule for Market Basket Data, Proceedings of the 1997 SIGMOD Conference on Management of Data, pp. 255-264, 1997.

[5] D. W. Cheung, J. Han, V. T. Ng and C. Y. Wong, Maintenance

of Discovered Association Rules in Large Databases: An Incremental Updating Technique, Proceedings of International Conference on Data Engineering, 1996.

[6] J. Han and Y. Fu, Discovery of Multi-Level Association Rules

from Large Databases. Proceedings of the 21st International Conference on Very Large Data Bases, pp. 420-431, Zurich, Switzerland, Sep. 1995.

[7] S. D. Lee, D. Cheung, B. Kao, A General Incremental

Technique For Maintaining Discovered Association Rules, Proceedings of the 5th International Conference On Database Systems For Advanced Applications, pp. 185-194, Melbourne, Australia, Apr. 1997.

[8] H. Ma nnila and H. Toivonen, Discovering Generalized

Episodes using Minimal Occurrences, Proceedings of the Second International Conference on Knowledge Discovery and Data Mining (KDD’96), pp. 146-151, Portland, 1996. [9] H. Mannila, H. Toivonen and A. I. Verkamo, Discovering

Frequent Episodes in Sequences, Proceedings of the First International Conference on Knowledge Discovery and Data Mining (KDD’95), pp. 210-215, Montreal, Canada, 1995. [10] J. S. Park, M. S. Chen, and P. S. Yu., Using a Hash-Based

Method with Transaction Trimming for Mining Association Rules, IEEE Transactions on Knowledge and Data Engineering, Vol. 9, No. 5, pp. 813-825, 1997.

[11] R. Srikant and R. Agrawal, Mining Sequential Patterns:

Generalizations and Performance Improvements, Advances in Database Technology–5th International Conference on Knowledge Discovery and Data Mining (KDD’95), pp.

269-274, Montreal, Canada, 1995.

[12] H. Toivonen, Discovery of Frequent Patterns in Large Data

Collections, Ph.D. thesis, University of Helsinki, Finland, 1996.

[13] M. J. Zaki, Fast Mining of Sequential Patterns in Very Large

Databases, Technical Report 668, The University of Rochester, New York, Nov. 1997.

历史街区保护与更新

历史街区保护与更新——以成都宽窄巷子为例 作者: 学号: 年级: 调查与分析 调研思路 调研报告的分析方法采用调查访谈与理论分析紧密结合的方法。调研过程有文献收集,确定调研方向,实地勘探、获取现状资料 、整理归纳资料 、分析结果。 调查主要采用下列方式、方法:现场勘踏、拍照、观测记录、数据统计分析、定量分析的问卷调查以及定性分析的访谈等,分析评价建立在坚实的数据资料基础上。问卷调查中总样本量为100份,有效问卷82份,有效率为82%;访谈中共10名被访者.在表达形式上,每一部分分析都由调查访谈的内容和理论知识两部分结合存在,充分体现了社会调研应联系实际,并最终分析解决问题。调查主要解决的问题是宽窄巷子是如何围绕休闲这一主题展开保护与改造,研究其规模、特色、空间处理以及如何与周边历史环境结合,侧重点是接到的尺度比例,如何处理使其更适宜现代人生活,在保护历史遗迹的同时,发展现代空间,不使之冲突。 【摘要】 当代历史建筑与街区保护对历史街区整体和区域的保护理念的强调,不仅强调对历史、艺术价值较高的文物性历史建筑进行原真性保护,而且主张对大量存在的一般性的历史建筑、街区以及存在期间的文化内容、生活方式进行整体的规划保护,同时也要求保持地区活力,瞒住现代生活的要求;重视对历史见证遗产的再利用,对于大量存在的一般性历史建筑既反对大拆大改,也不再主张不做任何变动的保护方式,而是主张保护和发展相结合,保护应到满足现代生活所需的内容。较好的协调保护与历史遗产与现代生活、传统保持与经济利益之间的矛盾。 【关键词】 街道尺度; 比例; 公共空间; 改造 前言 李白有诗云:“九天开出一成都,万户千门入画图”;在诗圣杜甫眼里,“晓看红湿处,花重锦官城”,诗人陆游又以“当年走马锦城西,曾为梅花醉似泥。二十里路香不断,青羊宫到浣花溪”的诗句,真切道出当年成都的优美环境与休闲生活。近年来,成都又先后荣获 调研流程图 深入研究阶段 验证阶段 结题阶段

历史文化街区保护管理办法

历史文化街区保护管理办法 历史文化街区保护管理办法历史文化名镇名村保护管理办法历史文化名城名镇名村保护规划编制办法 (征求意见稿) 2010年11月2日 历史文化街区保护管理办法 历史文化街区保护管理办法 目录 第一章总则 第二章历史文化街区的设立和撤销 第三章保护规划的编制、审批、公布、修改与备案第四章保护规划实施第五章监督管理 第六章附则 - 1 - 历史文化街区保护管理办法 第一章总则 第一条 (制定目的)为了加强历史文化街区的保护与管理,继承和弘扬中华民族优秀历史文化遗产,依据《中华人民共和国城乡规划法》、《中华人民共和国文物保护法》、《历史文化名城名镇名村保护条例》,制定本办法。 第二条 (历史文化街区的定义)本办法所称历史文化街区,是指经省、自治区、直辖市人民政府核定公布的保存文物特别丰富、历史建筑集中成片、能够较完整和真实地体现传统格局和历史风貌,并具有一定规模的区域。

第三条 (适用范围)历史文化街区的设立、规划制定与实施、监督和管理,适用本办法。 第四条 (管理主体层级)国务院建设行政主管部门会同国务院文物主管部门负责历史文化街区的管理工作。 地方各级人民政府负责本行政区域的历史文化街区的管理工作。 第五条 (保护原则)历史文化街区的保护应当遵循保护遗产,继承文化,科学规划,严格保护的原则,突出历史风貌的整体保护,保护历史遗存的真实性,维护风貌的完整性,维持功能的延续性,采取政府主导、居民参与的方式改善基础设施和居住环境,正确处理更新改造和保护的关系。 第六条 (保护资金)历史文化街区所在地的县级以上地方人民政府,根据本地实际情况安排保护资金,列入本级财政预算。 国家鼓励企业、事业单位、社会团体和个人参与历史文化街区的保护。 第七条 (支持奖励)县级以上地方人民政府及其有关部门对在历史文化街区保护工作中做出突出贡献的单位和个人,按照国家有关规定给予表彰和奖励。 第八条 (资质管理)承担历史文化街区保护规划的编制单位,应当符合相应的规划资质管理规定。 第二章历史文化街区的设立和撤销 第九条 (设立标准)历史文化街区必须具备下列条件: (一)有比较完整的历史格局和风貌; - 2 - 历史文化街区保护管理办法 (二)有比较丰富的历史建筑和历史环境要素; 2(三)占地面积不小于2.5hm;

深圳市城市更新办法实施细则解读

深圳市城市更新办法实施细则解读 1 2020年4月19日

《深圳市城市更新办法实施细则》解读 一、出台背景 ● 8月,广东省“三旧”改造政策出台,实现“三旧”改造用地出让、完善历史用地手续等六大突破。 ● 12月1日,《深圳市城市更新办法》正式施行,成为国内首部系统规范城市更新工作的政府规章,形成我市城市更新政策体系的核心。 ●明确了“政府引导、市场运作、规划统筹、节约集约、保障权益、公众参与”的原则,并确立了综合整治、功能改变和拆除重建三种更新模式。 ● 至今,市政府、市规划国土委相继出台一系列配套文件,形成了较为系统的城市更新政策体系,确立了“大力推进城中村和旧工业区改造、积极推进旧商业区改造、稳步推进旧住宅区改造”的策略,规范了相关运作流程。 ●《关于深入推进城市更新工作的意见》 ●《深圳市城市更新提速专项行动计划》 ●《深圳市城市更新单元规划制定计划申报指引》 2 2020年4月19日

●《城市更新单元规划审批操作规则》 ●《市规划国土委城市更新项目用地审查操作规程》 ●《深圳市城市更新项目保障性住房配建比例暂行规定》 ●《拆除重建类城市更新项目房地产证注销操作规则》 ●《深圳市宝安区、龙岗区、光明新区及坪山新区拆除重建类城市更新单元旧屋村范围认定办法》 ●…… ●当前,城市更新工作依然是市委、市政府高度重视的一项战略性工作。许勤市长在今年的政府工作报告中指出,“要以城市更新释放发展空间,推动城市发展由依赖增量土地向存量土地二次开发转变”,对城市更新工作寄予重望。●今年1月21日,市政府以 1号文的形式正式发布了《深圳市城市更新办法实施细则》。 二、主要原则及创新 ●原则一:与国务院《国有土地上房屋征收与补偿条例》、广东省“三旧”改造政策、市政府《深圳市城市更新办法》及《关于深入推进城市更新工作的意见》等相关文件保持衔接。 3 2020年4月19日

历史文化街区的保护和改造

浅谈历史文化街区的保护和改造 摘要:分析历史文化街区保护的意义及开发运作等问题。并结合孔城老街的实例,提出保护设想,希望找到既有利于文化保护和延续,又具有可操作性的改造方式。 关键词:历史街区、保护、开发 1引言 一座城市的特色,除了要有时代气息外,还要有历史文化传统。历史文化是城市文化遗产的重要组成部分,也是一个地区、一座城市悠久历史和灿烂文明的最好见证。同时,作为城市文化遗产保存最完整、最丰富的地区,历史文化街区被世代生活于此的人们倾注着很多复杂的情感。 2002年重新修订的《中华人民共和国文物保护法》明确规定历史文化街区所在地的县级以上地方人民政府应当组织编制专门的历史文化街区保护规划,并纳入城市总体规划。历史文化街区保护首次被以法律的形式确定下来。 2.历史文化街区保护的意义 2.2概念 历史文化街区是指经省、自治区、直辖市人民政府核定公布的保存文物特别丰富、历史建筑集中成片、能够较完整和真实地体现传统格局和历史风貌,并有一定规模的区域。《文物保护法》中对历史文化街区的界定是:法定保护的区域,学术上叫“历史地段”。 2.3意义 历史文化街区也是城市发展的重要资源,保护历史文化街区,对于建设特色城市有现实意义。文化被誉为"经济发展的原动力",这一点已经在很多城市的发展进程中得以证实。作为城市发展独特见证的历史文化街区,更是具有多方面的资源效应,在城市形象宣传、历史文化教育、乡土情结的维系、文化身份的认同、生态环境建设、和谐人居环境的构建等多方面具有综合的价值。 历史文化街区不仅仅是要保护构成历史风貌的文物古迹、历史建筑,还要保存构成整体风貌的所有要素,如道路、街巷、院墙、小桥、溪流、驳岸乃至古树等。历史文化街区是一个成片的地区,有大量居民在其间生活,是活态的文化遗产,有其特有的社区文化,不能只保护那些历史建筑的躯壳,还应该保存它承载的文化,保护非物质形态的内容,保存文化多样性。这就要维护社区传统,改善生活环境,促进地区经济活力。 3.历史文化街区的保护—以孔城老街为例 3.1历史沿革 孔城镇位于安徽省桐城市的东部,距今已有两千多年的历史。春秋中期(公元前770-481年),即为桐国南部门户,扼大别山东南区域通往长江水运通道之要冲,是古代重要的军事要塞,距今2500多年。三国(公元220-280年)时期,吴将吕蒙曾屯兵于此,建筑堡垒。距今1700多年。南宋时,孔城是民间抗金要塞,著名的抗金战场。元代为大别山东侧,江淮分水岭至长江大别山边缘地区的九大商镇之一。到明代有居民600余家,清代中、后期本镇发展为旧桐城县内“枞、汤、孔、练”四大名镇之首,达到古镇发展的高峰。现存约2公里长的石板古街和古建筑群就是晚清和民初遗留下来的历史文化遗产。 3.2老街概况 孔城老街是以南北向中大街为主轴,大街呈“S”形,地势南低北高,有主街一条,横街两条,另有七巷十三弄。连同两侧各甲之间(一甲到九甲)的巷道,形成鱼刺状的外部空间结构。总建筑面积约17万平方米。 街巷弄均为麻石所铺,店铺房舍皆青砖灰瓦,颇具江南水乡特色。著名的《桐乡书院四议》,戴钧衡即作于此,它为全国书院所奉崇,并载入《皇朝政典类纂》。镇东大沙河,平沙浩瀚,极目无垠,夕阳照射,宛如白雪,故有“孔城暮雪”之称。镇西南界荻埠河,碧波粼粼,

20150206上海市城市更新规划实施办法

上海市城市更新规划实施办法 (试行) (讨论稿) 目录 第一条 (目的) (1) 第二条 (定义和适用范围) (1) 第三条 (工作制度) (1) 第四条 (管理职责) (1) 第五条 (组织实施机构) (2) 第六条 (区域评估) (2) 第七条 (实施方案) (2) 第八条 (规划政策) (3) 第九条 (土地政策) (4) 第十条 (实施管理) (5) 第十一条 (公众参与) (5) 第十二条 (附则) (5) 附:上海市城市更新规划技术要求(试行)

第一条(目的) 为深化落实城市总体规划的要求,有效调动政府、市场、社会等多方资源,优化城市功能、完善公共设施,提升城市品质、改善城市环境、完善基础设施,激发城市发展活力,提高土地利用效率,有序推进城市更新,促进经济、社会、生态、文化协调发展,根据有关法律、法规,结合本市实际,制定本办法。 第二条(定义和适用范围) 本办法所称城市更新,是指为优化城市功能、完善公共设施,提升城市品质、改善城市环境、完善基础设施等而进行的各类建设活动。 本办法适用于在本市中心城、新城、新市镇已基本建成区域中认定的城市更新地区,已纳入土地征收范围的旧城改造、工业转型、城中村改造地区,应当按其它相关规定执行。 第三条(工作制度) 城市更新实行区域评估和实施方案相结合的工作制度,按照政府引导、市场运作、多方参与、利益共享的原则,依法合规地推进实施。 第四条(管理职责) 市规划土地管理部门负责统筹协调全市城市更新工作,依法制定规划和土地政策以及相关技术规范,指导、管理、监督区、县开展城市更新工作。 市政府其他相关职能部门应当在职责范围内,依法制定有关城市更新的配套政策和标准,依法履行相应的指导、管理和监督职责。

历史文化街区保护管理办法

历史文化街区保护管理办法 历史文化名镇名村保护管理办法 历史文化名城名镇名村保护规划编制办法 (征求意见稿) 2010年11月2日

历史文化街区保护管理办法 目录 第一章总则 第二章历史文化街区的设立和撤销 第三章保护规划的编制、审批、公布、修改与备案 第四章保护规划实施 第五章监督管理 第六章附则 第一章总则 第二章保护规划的审批、公布、修改与备案 第三章保护措施管理 第四章监督管理 第五章附则 第一章总则 第二章编制要求 第三章调查和评估 第四章历史文化名城保护规划编制 第五章历史文化街区保护规划编制 第六章历史文化名镇名村保护规划编制 第七章成果要求 第八章附则 附件一保护规划制图统一标准和要求

第一章总则 第一条(制定目的)为了加强历史文化街区的保护与管理,继承和弘扬中华民族优秀历史文化遗产,依据《中华人民共和国城乡规划法》、《中华人民共和国文物保护法》、《历史文化名城名镇名村保护条例》,制定本办法。 第二条(历史文化街区的定义)本办法所称历史文化街区,是指经省、自治区、直辖市人民政府核定公布的保存文物特别丰富、历史建筑集中成片、能够较完整和真实地体现传统格局和历史风貌,并具有一定规模的区域。 第三条(适用范围)历史文化街区的设立、规划制定与实施、监督和管理,适用本办法。 第四条(管理主体层级)国务院建设行政主管部门会同国务院文物主管部门负责历史文化街区的管理工作。 地方各级人民政府负责本行政区域的历史文化街区的管理工作。 第五条(保护原则)历史文化街区的保护应当遵循保护遗产,继承文化,科学规划,严格保护的原则,突出历史风貌的整体保护,保护历史遗存的真实性,维护风貌的完整性,维持功能的延续性,采取政府主导、居民参与的方式改善基础设施和居住环境,正确处理更新改造和保护的关系。 第六条(保护资金)历史文化街区所在地的县级以上地方人民政府,根据本地实际情况安排保护资金,列入本级财政预算。 国家鼓励企业、事业单位、社会团体和个人参与历史文化街区的保护。 第七条(支持奖励)县级以上地方人民政府及其有关部门对在历史文化街区保护工作中做出突出贡献的单位和个人,按照国家有关规定给予表彰和奖励。 第八条(资质管理)承担历史文化街区保护规划的编制单位,应当符合相应的规划资质管理规定。 第二章历史文化街区的设立和撤销 第九条(设立标准)历史文化街区必须具备下列条件: (一)有比较完整的历史格局和风貌;

城市更新改造项目合作协议书(范本)

城市更新改造项目合作协议书 协议编号: 甲方:股份合作公司 地址: 邮政编码:电话: 法定代表人: □身份证/□护照/□其它证件号码: 乙方: 地址: 邮政编码:电话: 法定代表人: □身份证/□护照/□其它证件号码: 根据《中华人民共和国土地管理法》、《中华人民共和国城市房地产管理法》、《中华人民共和国合同法》、《深圳市城市更新办法》及相关法律法规、政 策规定,为改变拟更新改造区域[1]----片区老工业厂房/旧住宅区城市基础设施、公共服务设施配套差/产业收益低/环境落后/存在安全隐患/土地用途和利用效率与规划功能不符等状况,甲乙双方经多次协商,本着“改造增值,优于现状收益,互惠互利共赢”的原则,就------片区的城市更新项目的合作事宜订立合同,内容如下,双方共同遵守履行。 第一条项目更新范围及现状 本城市更新改造项目获得相关部门批准纳入城市更新单元规划制定计划后, 项目区域现状见附件一。(本城市更新改造项目的城市更新单元规划制定计 划的申报材料:地块现状一览表、更新单元范围图、现状权属图、建筑物信 息图。) 本城市更新改造项目的专项规划情况见附件二(城市更新单元规划通过审批,可附本城市更新项目的专项规划) 甲方在项目区域内的房地产占地(包括构筑附属物)平方米,建筑面 积平方米(其中住宅面积平方米,厂房面积平方米),详细情 况见附件三。 第二条项目改造期限: 2.1本协议签订后,应在个月内(建议12个月内)制定房地产收购搬迁补 偿安置方案,并通过甲方的股东代表大会表决同意; 2.2本协议签订后,应在个月内(建议24个月内)完成%以上(建议90%以上)的房地产收购搬迁补偿安置协议的签订工作;分期建设的,应在个月

历史文化街区国内外研究现状

历史文化街区国内外研究现状 我国对历史文化街区的研究始于1986年,但直到2012年,有关历史文化街区的研究文献数量才开始激增。尤其随着历史文化街区保护实践的不断深人,学术界对历史文化街区的研究不断跟进,形成了不少有价值的研究成果。归纳起来,学者们对历史文化街区的研究主要集中在以下几个方面。 1.历史文化街区的概念研究。 目前,学术界对“历史文化街区”相关概念的表述有多种称谓,如“历史地区”、“历史地段”和“历史城区”等。“历史文化街区”作为2002年以后我国历史文化名城保护体系中“中观层面”的核心概念,其前身为1986年提出的“历史文化保护区”。2002年10月修订的《中华人民共和国文物保护法》规定,所谓“历史(文化)街区”是指保存文物特别丰富并具有重大历史价值或者革命意义的城镇、街道、村庄。国家建设部和国家文物局起草的《历史文化名城保护条例(送审稿)》中,将“历史文化街区”定义为城市中保留遗存较为丰富,能够比较完整真实地反映一定历史时期传统风貌或民族地方特色,存有较多文物古迹、近现代史迹和历史建筑,并具有一定规模的地区。其中,前者指代了城镇和乡村环境中的特定区域,而后者则特指城镇环境中的某一特定区域,且前者的规模大于后者。2012年,《历史文化名城保护规划规范》(GB50357—2012)的发布,使“历史城区”、“历史地段”、“历史文化街区”等成为我国名城保护体系“中观层面的规范用语。其中,“历史地区”的范畴最广、包含内容最多,其他概念都内含于“历史地区”之中;“历史文化街区”又内含于“历史地段”之中,两者的主要区别在于前者为“经省、自治区、直辖市人民政府核定公布应予重点保护的历史地段。 2.历史文化街区保护与更新研究。 (1)历史文化街区的保护现状与存在问题。 近年来,随着城市化进程的加快,历史文化街区保护问题受到越来越多的学者的关注,尤其对历史文化街区开发中暴露出的诸多问题,学者们给予了更多

武汉市历史文化风貌街区保护规划和修建性详细规划编制技术规定分析

武汉市历史文化风貌街区保护规划和修建性 详细规划编制技术规定(试行) 时间:2014年03月05日来源:本站原创作者:武汉魅力村镇点击:522次 1.1依据为了保护历史文化资源,规范历史文化风貌街区保护规划的编制和管理,根据《中华人民 共和国城乡规划法》、《中华人民共和国文物保护法》、《城市紫线管理办法》、《历史文化名城保护规划规范(GB50357-2005) 1.1 依据 为了保护历史文化资源,规范历史文化风貌街区保护规划的编制和管理,根据《中华人民共和国城乡规划法》、《中华人民共和国文物保护法》、《城市紫线管理办法》、《历史文化名城保护规划规范(GB 50357-2005)》、《武汉市旧城风貌区和优秀历史建筑保护管理办法》(2003年)、《武汉市历史文化名城和优秀历史建筑保护条例》(2013),结合武汉市历史文化资源保 护和规划管理的实际需要,制定本技术规定。 编制历史文化风貌街区保护规划及修建性详细规划,应依据《武汉市历史文化名城保护专 项规划》、《武汉主城历史文化与风貌街区体系规划》、《武汉市都市发展区紫线专项规划》、《武汉市主城区紫线专项规划》、《武汉市主城区控制性详细规划导则》等上位规划及本技术规定进行编制。 1.2 作用 本技术规定是结合武汉市历史保护的特点,对《历史文化名城保护规划规范 (GB 50357-2005)》中历史文化街区保护规划编制内容的深化。编制完成的保护规划和修建 性详细规划将作为武汉历史文化风貌街区建设和管理的依据。 1.3 适用范围 本技术规定适用于武汉市主城区历史文化风貌街区(包括历史文化街区、历史地段及传统 特色街区)保护规划和修建性详细规划的编制。其他历史文化资源集中的区域可参照本技术 规定执行。 1.4 规划编制组织、审批、修改程序 市规划主管部门应当会同市房屋、文物主管部门组织编制历史文化风貌街区保护规划,并征求所在地区人民政府和相关管理部门的意见,依法报市人民政府批准后向社会公布。 市规划主管部门应当根据历史文化风貌街区保护规划组织编制历史文化风貌街区修建性 详细规划。

历史文化街区与历史建筑的保护

历史文化街区与历史建筑的保护——以开封市双龙巷历史文化街区的保护规划为例 历史文化街区是历史文化名城的重要组成部分,也是名城历史文化和风貌的体现,而其中的历史建筑是构成历史文化街区的主体,历史建筑的数量也是确定历史文化街区最重要的依据. 对于历史建筑的保护研究目前还处于起步阶段,本文以开封市双龙巷历史文化街区的保护规划为例,对历史建筑的保护与利用提出建议. 对历史建筑的保护措施以修缮和改善为主,即不改变其外部历史风貌,对其内部结构,空间布局,内部设施,使用功能做适当的变动,以达到永续利用的目的,鼓励有利于保护的各种使用功能,引用市场机制解决历史建筑的保护、维修、臵换等,探索出适用于历史建筑的保护新路. 项目概况 开封市双龙巷历史文化街区位于开封市老城区东偏北的区域,是老城区内划定的两片历史文化街区之一,相邻周边用地多为未经改造的低层民居,建筑风格及色彩,空间尺度上较为协调,保持着良好的古城风韵.双龙巷传统民居保护区以北方古城传统四合院住宅形式为主,是清末民初民用建筑的典范,较好地保留了当时的居住格局,在开封历史文化名城保护中具有相当重要地位.

居住在双龙巷的居民几乎人人皆知双龙巷的故事,相传宋代皇帝赵匡胤, 匡义两兄弟曾住该巷,古代皇帝称真龙天子,故得名后人在双龙巷西口两侧各筑一龙头以示纪念.后开封城几遭破坏,现在仅存在的一个龙头为清代所建,其表面虽已斑驳,却依然是附近居民的骄傲. 保护范围及价值 双龙巷历史文化街区的保护范围是域,其核心保护区范围北至水车胡同,轱辘湾街一线,南至朝阳胡同,东至塘坊口街,南聚奎街,屈家胡同一线,西护面积16l7公顷.建设控制区范围北至以开封市双龙巷历史文化街区的保护规划为例延寿寺街,南至财政厅东街,平等街一线,东至内环东路,西至北道门,右司官口,北兴街一线,面积侣.8 公顷. 双龙巷历史文化街区的价值是以双龙巷、双井巷为代表的开封古城内保存最完整、规模最大的传统居住街区. 历史建筑集中,风貌相对完整.留存较多的有传统特色和文革历史特色的民居院落,有关于赵匡胤、史可法等传说。规划要点搬迁区内所有工厂,为街区保护提供发展用地.恢复僧忠亲王祠、史可法祠等历史古迹.最大范围的保护双龙巷街区历史风貌完整的区域,积极修缮历史建筑,改善居住环境.划定有价值的历史建筑为保护建筑,重点修缮,并整治建设周边环境,改造基础设施,尤其道路改造及宅院内部给排水工程改造. 公认的三个历史街区的核定标准:历史真实性、生活真实

城市更新项目操作程序

城市更新项目操作程序

————————————————————————————————作者:————————————————————————————————日期:

城市更新项目操作程序 城市更新单元规划计划 受理部门大致工作及主要审批程序所需材料最终结果 辖区街道办1、申报主体与权利人签订合作开发及相关协议; 2、按政府相关规定多个或单一权利人委托一个公司或个人做为申报主体; 3、权利人或申报主体向街道办提出申请,要求在划定范围纳入城市更新计 划;主要审核权利人的更新意愿及土地权属的真实情况; 3、提交《指引》规定的申报资料预审。 1、拆迁范围内房屋权利人或申报主体致函所辖街道办; 2、提供上述文字图纸及电子版本的相关资料。 街道办审核后致函区旧改办,阐 述街道办对更新单元的意见。 区旧改办1、区旧改办审核更新单元范 围内的建筑物的权属是否清晰;权利人的更新意愿是否符合要求;满足上述 要求,核发复函; 2、将区旧改办的复函连同相 关申报资料报区国土局更新办,核实更新范围内建筑物产权及土地权属情 况;核实该区域是否与法定图则规定相冲突;无异议则给予复函;并转呈市 规土委更新办审批; 1、初步划定拆除重建区域作为城市更新单元范围,并 将该区域内地上建筑物、构筑物逐一拍摄照片,统 计拆除建筑面积等相关数据; 2、委托设计院将初步划定的更新单元范围依据《深圳 市城市更新单元规划制定计划申报指引(试行)》之 规定绘制范围图、现状权属图、建筑物信息图及更 新单元规划制定计划申报表、更新单元规划制定计 划申请书、地块现状详细信息一览表、更新意愿表 (包括上述信息的电子光盘)。 审核并确认申报主体。 区旧改办收到辖区街道办同意更新改造的函后,申报主体将《指引》规定的资料呈报 给旧改办。 旧改办主要审核: 1、审核申报区域权利主体权属情况是否清晰; 2、该区域是否与法定图则规定相冲突; 3、该区域是否与其他宗地红线重叠; 4、单一权利主体的申报,区旧改办无需做更新意愿的公示程序; 5、多个权利主体申报纳入更新单元规划计划的,必须做更新意愿的公示 15个自然日。 辖区街道办的批复,同意该更新单元纳入更新单元规划 制定计划及上述文字资料。 区旧改办汇总公示意见呈区规 划国土局更新科。 区国土局更新科旧改办将公示的意见整理汇总并提交给区规土局更新科。 1、区规土局更新科牵头征求各职能部门意见并会签意见后,提交区规土局 上会审议、批复; 2、市规土委相关部门组织申报主体单位共同委托指定有资质设计院拟订 城市更新单元规划草案。 草案提交区国土局上会审议并公示; 市城市更新办征求相关部门意见后,45个工作日内形成 审定的规划草案,报请市规划国土委技术委员会审议。 汇总公示后的意见。 上报市规委 区旧改办区旧改办或区规土局牵头组织申报主体委托有资质的设计单位,对更新区凭区规土局同意纳入更新单元规划制定计划的批复。《城市更新单元规划制定计划

浅析传统街区街道空间

城市设计方法与理论·课程论文·2013年6月 浅析传统街区街道空间 姓名 (学校) 摘要:传统街区以独特的空间环境形态和场所魅力焕发起人们的思考。分析研究传统空间的环境形态及其特征,探讨传统街区场所营造的城市设计原则, 是传统街区改造更新的基本方法和途径。研究了传统街区的街道空间,并对街巷的平面形态、街巷的比例与尺度、街巷的结点及其结点的形态进行了分析,指出充分了解街道空间对传统街区的物质文化和非物质文化具有重要意义。 关键词:传统街区建筑空间场所营造设计理念 1引言 传统街区带着信息的建筑,蕴含着改变人和历史的力量。这些衰败的建筑从单点的保存,延伸至再生使用,更延伸到周围新建筑在设计时也会考虑与它们配合,这样就从点连成线,再发展成面,使地方呈现更和谐的面貌。而它们之所以能扩展,也是因为再生或设计的着眼点都在于如何使生活更好和环境更好。传统街区是指具有城市发展历史文化载体的城市区域,由具有历史文化特征的建筑群,传统街道以及广场空间组成,反映了历史文化和城市的特色。传统街区应当具有城市历史风貌的相对完整性而且应是具有真实生活性的街区,我国传统街区的空间形态明显受到了传统文化的影响,特别是古代哲学、政治、社会伦理等因素,传统街区的空间形态按照其表现和感知形式我们可以将其分为显性的物质空间环境要素和隐性的社会文化要素两部分。下面由传统空间形态规划的特征进行分析: 2特有的空间组合观念 从我国古代的城市和建筑的布局中可以看出都非常地重视建筑之间的空间院落的组合变化$建筑组合一般表现为多层次多院落和多变化,一个院落接着一个院落从而形成不断发展下去的空间组合形式,从我国不同地区的同性质的民居的比较来看,如北京的四合院、山西晋中一带的窄形多进深住宅以及安徽徽州地区的民宅等多是在建筑空间的组合上追求叠屏封闭曲折幽静小中见大的空间环境,从这里我们可以看出我国传统城市中普遍采用的院落组织的空间形式是传统生活个性观念的强烈反映。 3传统街区空间构成要素 传统街区空间的组成形式是多种多样的在不同的地域不同的文化习俗下有着自己的地域文化特色,就其传统街区空间构成要素依其表现形式和感知的空间形态可以划分为显性的物质空间环境要素和隐性的社会文化要素两种表现形式,物质空间要素是街区空间构成的基础!但是社会文化对传统街区空间的影响也是不容忽略的显性的物质环境要素街巷空间在传统街区空间中!街巷空间占有极其重要的地位!可以说街巷是传统街区的骨架,决定着街区形态的大致结构,传统街巷是在街区的发展过程中逐渐形成的,是建筑的衍生是由周边建筑所围合而成的"街巷空间的丰富的变化,正是由于建筑的自发建造。才形成了如此丰富多变的形式。 4传统街区空间内涵与特征 传统街区是经历了一个相对漫长的历史过程形成的人类生活聚居的场所。它的空间环境承载着

历史街区的保护与规划专题研究

历史街区的保护与规划专题研究 摘要:本文是对历史街区的保护与规划专题的总结性思考。结合这一学期所学,对历史街区的保护与规划专题的认识,汇报案例(绍兴仓桥直街规划)的介绍和总结,对本专题同学所做的内容的总结和思考,以及对历史街区保护与规划的个人观点。 关键词:城市建设历史街区保护规划文化 随着对历史街区问题的研究不断深入,城市建设与历史街区的保护及规划的问题已不仅仅是规划师或建筑师的研究课题,更关系到市民的切身利益,保护的内容也不再是某一文物,某一街区的形态特征,在某种意义上是使城市生活文化得以更好延续。历史街区保护是一项具有社会实践性特征的复杂工作,既需要不断推进城市保护理念的提升,还需要不断加强保护规划的实施以及控制机制的完善。历史文化遗产凝结了民族的智慧,对我们现代城市的硬件和人们文化思想的软件发展有极其重要的指导作用,然而,目前看来,历史文化遗产与城市建设的矛盾还没有得到很好的解决,本文拟就城市建设与历史街区的保护及规划的问题研究。现行的历史街区保护规划中存在着一些问题,这些问题的根源主要体现在经济、文化及保护制度上,通过研究历史街区的本质特征,对历史街区的保护规划实践进行思考,以总结城市建设过程中历史街区的保护及规划的经验和教训。 城市建设与历史街区的保护及规划的问题研究主要概括为三个方向,与时代的关系,与自然的关系,与文化的关系。与时代的关系:要正确认识到时代变迁是引发历史街区保护矛盾的根源,在保护过程中强调分析,针对不同类型寻求不同策略。探寻历史街区发展源动力,寻求新的社会因素的介入。与自然的关系:人居环境的建设可以追溯到整个人类进化以及迁徙历程,人类对于生存环境的选择受制于当时的生产水平,反映了人类改造自然寻求理想人居环境的历程。与文化的关系:每一个国家、民族和地区都有自己独特的气候条件、地理条件、文化传统、生活习惯和建筑风格,城市开发需要对历史文脉和地方文脉的重视和恰当的运用。在城市开发和建筑设计中,引用国际城市规划理论成果是必要的,但是如果抛弃了自己本土文化和悠久历史所沉淀下来的建筑经验和理论,就是本末倒置了。另外,对待传统街区的复原也要持审慎的态度,传统的美不在于符号的简单堆砌,而表现在经过艺术考虑的内在结构。建筑物的营造如同写文章,应从实际需求出发,理清地区文化脉络,将文化符号、地域符号进行重组。城市是我们赖以生存的家园,保护和发展历史街区,进行合理的规划安排,结合时代,自然,文化的关系,使历史街区在保护中发挥它的重要价值。 结合本专题研究,找的案例为浙江省绍兴仓桥直街。绍兴仓桥直街全长一点五公里的绍兴仓桥直街由河道、民居、街坊三部分组成,民居多为清末民初建筑,众多富有地方特色的台门保存完好,具有浓郁的水乡风貌。2003年获“二零零三年联合国教科文组织亚太地区文化遗产保护优秀奖”。 绍兴仓桥历史街区保护与整治是国内成功的保护规划,而在规划设计上采用

将军庙历史文化街区保护规划

将军庙历史文化街区保护规划 一、规划范围 南至泉城路,北临大明湖路,东至鞭指巷,西临太平寺街、西城根街,总用地面积16.08公顷。 图1 现状航拍图 二、规划原则

以科学发展观为指导,遵循保护遗产本体及环境的真实性、完整性、生活延续性和保护利用的可持续性的原则,保护历史文化遗产,引导商业发展,改善人居环境,促进经济社会协调发展。 三、指导思想 规划坚持保护为主、合理利用、改善环境、加强引导,有效管理的指导思想,按照“政府主导,居民参与,实体运作,小规模、渐进式更新”指导思路开展工作。 四、规划目标 1、保护街区整体空间格局和历史风貌,尤其是体现街区价值特色的传统街区风貌,包括:具有济南“城泉共生”特色的传统居住区;济南古城多元宗教信仰的文化中心;多类型建筑集中展示区;济南古城内中西文化交汇融合、商贸交流繁荣之地。 2、保护街区丰富的历史文化遗存,保持历史文化街区的传统居住功能为主,使其成为城市发展历史的见证。 3、整治街区环境,完善基础设施配套,优化人居环境,促进社区健康发展,使其成为传统居住文化突出的特色社区。 4、结合街区现状情况,综合研究符合街区特点的规划实施模式,提出规划实施建议,加强保护规划中提出的对街区保护与整治措施的可行性。 五、街区价值与特色 作为济南古城现存的保留较完整的、最具传统特色的地区,将军庙街区具有历史、艺术、科学研究方面的极高价值。街区是古城的传统居住区;街区还集中了大量宗教建筑,是中西

文化交流繁荣的区域;街区内整体风貌保存尚好,传统街巷空间基本保留完好,名泉水体、大树丰富,建筑遗产多样。将军庙街区的总体价值为: 1、将军庙街区是具有济南“城泉共生”特色的传统居住区之一; 2、将军庙街区是济南古城宗教建筑分布最为集中,同时最能体现多元宗教信仰的文化中心; 3、将军庙街区建筑类型多样,体现了中西建筑精湛的建造技艺; 4、将军庙街区是济南古城内中西文化交汇融合、商贸交流繁荣之地。 六、保护内容及保护要素 本规划从物质文化遗存要素和非物质文化遗存要素两部分确定保护内容及各项保护要素。其中物质文化遗存要素包括传统格局、传统街巷、不可移动文物、历史建筑、泉水历史文化遗存、历史环境要素等七个方面;非物质文化遗存要素包括非物质文化遗产及优秀传统文化要素和与泉水相关的优秀传统文化要素。

历史街区保护与更新常见模式及案例分析

历史街区保护与更新常见模式及案例分析 摘要:历史街区的保护是当下很多人关心的问题,过快的城市更新速度已经严重的威胁到了它们的存在。本文对当下的一些历史街区保护模式进行了总结,并分别分析了其成功的方式,希望可以为相关部门提供一点参考。 关键词:历史街区;保护;更新 Abstract: the historical block protection is instantly many people care about, and the rapid urban renewal rate has a serious threat to their existence. In this paper, the present some historical block protection mode are summarized, and analyzes its successful respectively way, the hope can provide some reference for the related department. Key words: the historical block; Protection; update 历史是一面镜子,我国是四大文明古国之一,历史尤为悠久,很多古老的建筑既见证着我们曾经的辉煌,又具有很高的研究价值,因此必须予以积极保护。但是近年来,随着我国城市化运动的不断高涨,原有的建筑逐步被高楼大厦所取代,这其中有很多历史建筑也惨遭不幸,造成了非常大的经济损失和人文损失,有志之士已经注意到了这一点,积极呼吁相关部门的保护和支持,也取得了一定的成效,但是这项工作任重而道远,需要我们所有人都有足够的认识才是根本之道。 很多古城的历史文化气息浓厚,古建筑非常多,难能可贵的是,这些建筑都是成片存在的,我们称之为历史街区,这些历史街区的存在让我们对过去的生活有了很多的理解,价值极大,但是只是单纯的进行保护效果又不好,最好的方式就是保护与利用合为一体,使其重换青春,本文即基于这样的目的做了探讨,希望能引起相关部门的重视。 我国的历史街区归纳而言有以下四种,其一是以古城文化为依托发展文化经济的历史街区,如丽江、平遥等均通过政府资金进行了大规模的整伤修缮,成为了著名的文化旅游城市。其二是处于城市黄金地段的老商业街区和办公街区的“置换”,典型案例就是上海的外滩。其三是以艺术家自发改造为主的产业建筑历史街区,最初是在上海,一些艺术家在苏州河沿岸地区自发改造的中小型产业建筑遗产成为了艺术家的工作室,随后在北京、杭州、成都出现了有类似的项目。其四是城市居住型历史街区的商业化改造,最典型的就是上海的“新天地”,随后这一模式迅速在全国范围内推广。

江苏省历史文化街区文化街区保护规划编制导则(试行)

江苏省历史文化街区 保护规划编制导则(试行) ─1─江苏省建设厅文件苏建规〔2008〕110号 关于印发《江苏省历史文化街区保护规划 编制导则(试行)》的通知 各市、县规划局、建设局,各城市规划编制单位: 为加强对全省历史文化街区保护规划编制工作的指导,提高全省历史文化街区保护规划工作的质量和水平,我厅组织起草了《江苏省历史文化街区保护规划编制导则(试行)》。 现印发给你们并从2008年6月1日起试行,请结合当地的实际组织实施。试行过程中如有意见和建议,请反馈给我厅城乡规划处。 联系人:方芳 联系电话:025-********(兼传真) 江苏省建设厅二○○八年四月八日 ─1─目录 一、总则 1.目的 2.适用范围 3.规划编制的依据 4.规划编制的原则 5.规划的层次定位 二、保护规划编制管理

1.规划编制组织 2.编制单位的资质 3.规划审批 4.规划修编 5.公众参与 三、保护规划文本大纲内容 1.规划总则 2.历史文化遗存的保护利用 3.用地、人口与空间规划 ─2─ 4.交通与市政设施规划 5.技术指标与经济测算 6.实施政策建议 四、保护规划成果内容与要求 1.保护规划成果内容 2.保护规划成果要求 五、保护规划技术指引 1.历史研究 2.现状调查与评估 3.历史文化遗存保护利用 4.用地、人口与空间规划

5.交通与市政设施规划 6.技术指标与经济测算 7.规划实施政策建议 ─3─ 江苏省历史文化街区保护规划编制导则 (试行) 一、总则 1.目的 为了加强对全省历史文化街区保护规划编制的指导,规范历史文化街区保护规划编制工作,根据相关法律、法规、规范和文件,结合我省实际,制定本导则。 2.适用范围 本导则适用于我省历史文化名城、名镇中的历史文化街区保护规划的编制及其管理工作。历史文化保护区和历史文化名村保护规划的编制可参照执行。 3.规划编制的依据 3.1《中华人民共和国城乡规划法》、《中华人民共和国文物保护法》等相关法律,《江苏省历史文化名城名镇保护条例》等法规; 3.2《城市紫线管理规定》等部门规章; ─1─3.3《历史文化名城保护规划规范》等相关技术规范、标准; 3.4《省政府办公厅转发省建设厅省文物局关于加强历史街区保护工作意见的通知》(苏政办发〔2007〕86号)等相关规范性文件; 3.5城镇总体规划,历史文化名城、名镇保护规划以及相关上位规划。

上海城市更新实施细则(试行)

附件1 上海市城市更新规划土地实施细则(试行) (征求意见稿) 第一章总则 第一条(目的) 为实施《上海市城市更新实施办法》(以下简称“《办法》”),规范本市城市更新活动,建立科学、有序的城市更新实施机制,制定本细则。 第二条(工作原则) 城市更新应当坚持以下原则: 规划引领,有序推进。落实区域评估、整体更新的要求,发挥规划的引领作用,依法推进试点,实现动态、可持续的有机更新。 注重品质,公共优先。坚持以人为本,激发都市活力,提升城市品质和功能,优先保障公共要素,改善人居环境,增强城市魅力。 多方参与,共建共享。搭建实施平台,创新规划土地政策,使多元主体、社会公众、多领域专业人士共同参与,实现多方共赢。 第三条(城市更新工作领导小组) 城市更新工作领导小组由市政府及相关管理部门组成,

负责领导全市城市更新工作,对全市城市更新工作涉及的重大事项进行决策。城市更新工作领导小组下设办公室,办公室设在市规划和国土资源主管部门,负责制定工作机制,统筹协调相关部门工作,完善配套政策。 第四条(市规划和国土资源主管部门的职责) 市规划和国土资源主管部门应当根据《办法》及本细则的规定负责城市更新的日常管理工作,包括制定技术规范、管理规程,组织、协调和监督全市的城市更新工作,指导城市更新的实施,开展城市更新政策的宣传工作。 第五条(区县人民政府的职责) 区县人民政府应当根据《办法》及本细则的规定,负责本辖区内的城市更新工作,包括制定城市更新区域评估报告和实施计划;指定相应部门作为专门的组织实施机构,协调本辖区内的城市更新工作;定期评估城市更新项目的实施情况,督促更新项目依法合规并按计划落实。 第二章城市更新区域评估 第六条(区域评估的工作要求) 区县规划和土地管理部门应会同区县相关部门、街道办事处或镇政府组织开展区域评估,确定地区更新需求,进行地区公共要素评估,划定城市更新单元(以下简称“更新单元”),明确适用更新政策的范围和要求,形成区域评估报告报审稿。

传统村落街巷空间的保护与更新_以金丘村_民俗历史文化展示区_为例_沈一凡

传统村落街巷空间的保护与更新 ——以金丘村“民俗历史文化展示区”为例 杭州中联程泰宁建筑设计研究院沈一凡 景宁畲族自治县规划设计室柳晓杰 浙江理工大学建筑工程学院洪艳 摘要:传统的街巷空间是一个传统村落的空间精华所在,也是村落风貌的重要组成部分。如果把村落视作一个肌体,那么街巷无疑就是它的血脉。随着科技的发展、现代城市规划理论的出现,传统街巷的设计变得越来越理性,现代生活的需要使其布局、交通、管线等物质功能被强化。然而,街巷不仅仅是物质的载体.它还是人们生活的场所和舞台,承载着人们的生活和记忆。 关键词:传统村落街巷空间;传统建筑;保护与更新;文化传承 Abstract:The traditional street space is the space of a traditional village in essence, is an important part of village character. If the village as a body, then the street is no doubt that it's blood. With the development of technology, the emergence of modern urban planning theory, the traditional street design become more rational, modern life need to layout, traffic, pipelines and other physical function is strengthened. However, the street is not only the carrier substance. It is a place people live and the stage, carrying the people's life and memory. Key words:historical and cultural streets; traditional building; protecting and updating 1.金丘村“民俗历史文化展示区”保护与更新的背景 1.1 畲族传统文化慢慢流逝 金丘村村民缺乏民族文化的自我保护意识。在汉族文化的强势冲击下,畲民们看到的恰恰是强势的汉族文化和与之对应的物质利益,加上对强势文化的崇尚心理,于是出现盲目趋向外来文化的现象。 ·468·

相关文档